**Genetic & Evolutionary Biometrics**

Aniesha Alford, Joseph Shelton, Joshua Adams, Derrick LeFlore, Michael Payne, Jonathan Turner, Vincent McLean, Robert Benson, Gerry Dozier, Kelvin Bryant and John Kelly

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/51386

## **1. Introduction**

Genetic & Evolutionary Computation (GEC) is the field of study devoted to the design, development, and analysis of problem solvers based on natural selection [1-4] and has been successfully applied to a wide range of complex, real world optimization problems in the areas of robotics [5], scheduling [6], music generation [7], aircraft design [1], and cyber secur‐ ity [8-11], just to name a few. Genetic and Evolutionary Computations (referred to as GECs) differ from most traditional problems solvers in that they are stochastic methods that evolve a population of candidate solutions (CSs) rather than just operating on a single CS. Due to the evolutionary nature of GECs, they are able to discover a wide variety of novel solutions to a particular problem at hand – solutions that radically differ from those developed by tradition‐ al problem solvers [3,12,13].

GECs are general-purpose problem solvers [1,2,4]. Because of this fact and their ability to hybridize well with traditional problem solvers [1], a number of new subfields have emerged. In the field of Evolutionary Robotics [5,14], GECs are used in path planning [15], robot behavior design [16], and robot gait design [17]. In the field of Evolutionary Design [1,18], GECs are being used to evolve lunar habitats [19], emoticons [20], and music [7,21]. GECs have also been used successfully in a wide variety of scheduling applications [22,23] – which in turn has spawn a subfield known as Evolutionary Scheduling [6,24].

Currently we are seeing the emergence of a new and exciting field of study devoted to‐ wards the design, development, analysis, and application of GECs to problems within the area of biometrics [25-29]. We refer to this new subfield of study as Genetic and Evolutionary Biometrics (GEB) [25-27,31]. In this chapter, we will provide a brief history of GEB as well as

introduce a number of GEB applications (which we refer to as GEBAs). The GEBAs present‐ ed in this chapter are actually hybridized forms of traditional methods used within the bio‐ metrics research community. These GEBAs typically evolve solutions that are radically different from and, in many instances, more efficient than solutions developed by traditional meth‐ ods currently used within the biometrics research community.

**2.2. A Brief History of GEB**

of the system.

tures than the GA.

extraction [26-29], and biometric-based security [9-11,39].

*2.2.1. GEBAs for Feature Selection and Weighting*

Within the field of biometrics, GEB applications (GEBAs) have been successfully applied to the areas of feature selection [25,30,31,33,35,36], feature weighting [25,30,31,37,38], feature

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386 103

Concerning GEBAs for feature selection, Galbally et al. [35] developed Genetic Algorithms (GAs) [1] for feature selection and applied them in an effort to solve the signature verifica‐ tion problem. The GAs were applied to two training sets which were formed using the sig‐ natures of 330 subjects from the MCYT Signature database [40]. The first training set consisted of five signatures of each subject and the second training set consisted of 20 signa‐ tures of each subject. A test set was formed using the remaining signatures. Their results showed that the GAs, when compared to the baseline method that used all of the features, were able to reduce the number of features used while improving the recognition accuracy

Ramadan and Abdel-Kader [36] applied two GECs, Particle Swarm Optimization (PSO) [13] and a GA, for feature selection for a facial recognition problem. Images from the Cambridge ORL database [41] were used to form a training set, which consisted of four images of each of the 40 subjects, and a test set, which consisted of six images of each of the 40 subjects. Two baseline methods, the Discrete Cosine Transform (DCT) [42] and Discrete Wavelet Transform (DWT) [43] methods, were then used to extract the original set of features. Rama‐ dan and Abdel-Kader demonstrated that their GECs could increase the recognition accuracy over the baseline methods while using fewer features. In addition, the PSO used fewer fea‐

Kumar et al. [44] also applied two GECs, a Memetic Algorithm (MA) [45] and a GA, for feature selection for a face recognition system. The GECs were tested on two facial databases: the Cambridge ORL database [41] and a subset of 20 subjects from the YaleB database [46]. The following feature extraction techniques were used to create the original feature templates: Principal Component Analysis (PCA) [47], Linear Discriminant Analysis [48], and Kernel PCA [49]. The MA and GA were applied in an effort to reduce the dimensionality of the feature templates as well as to increase the recognition rate. Their results showed the GECs outper‐ formed the baseline methods, which used all of the extracted features, in terms of feature usage and recognition rate. Their results also showed that the MA outperformed the GA.

In [37], Abegaz et al. applied two GEBAs for feature selection and feature weighting for fa‐ cial recognition. The first GEBA, referred to as Genetic & Evolutionary Feature Selection (GEFeS), evolved subsets of biometric features also in an effort to increase the recognition rate while reducing the number of features used. The second GEBA, referred to as Genetic & Evolutionary Feature Weighting (GEFeW), evolved weights for the features. These GEBAs, which were instances of a Steady-State Genetic Algorithm (SSGA) [1-4] and Estimation of Distribution Algorithms (EDAs) [50], were applied to Eigenface[51] and LBP facial feature templates. Their results showed that GEFeS and GEFeW outperformed the baseline meth‐

The remainder of this chapter is as follows. In Section 2, we discussGECs and provide a brief history of the field of GEB. We also present an overview of the Local Binary Patterns (LBP) [32-34] method and in Section 3 we present the GEBAs used in this work. In Section 4, we describe the experiments performed and we present our results. In Section5,we provide an additional discussion of our results, and in Section 6, we provide a summary of our work as well as directions for further research.

## **2. Background**

#### **2.1. Genetic & Evolutionary Computation (GEC)**

GECs typically work in the following fashion [1-4,12,13]. Initially, a population of CSs is randomly generated. Each CS is then evaluated and assigned a fitness based on a user-de‐ fined evaluation function. Next, parents are selected from the population based on their fitness and are allowed to produce offspring CSs. The offspring are then assigned a fitness and usually replace the worst performing CSs within the current population. This evolutionary process is then repeated until a user-specified stopping condition is satisfied. Figure 1 provides a flow‐ chart of the typical GEC process.

**Figure 1.** Flowchart of a typical GEC.

## **2.2. A Brief History of GEB**

introduce a number of GEB applications (which we refer to as GEBAs). The GEBAs present‐ ed in this chapter are actually hybridized forms of traditional methods used within the bio‐ metrics research community. These GEBAs typically evolve solutions that are radically different from and, in many instances, more efficient than solutions developed by traditional meth‐

The remainder of this chapter is as follows. In Section 2, we discussGECs and provide a brief history of the field of GEB. We also present an overview of the Local Binary Patterns (LBP) [32-34] method and in Section 3 we present the GEBAs used in this work. In Section 4, we describe the experiments performed and we present our results. In Section5,we provide an additional discussion of our results, and in Section 6, we provide a summary of our work as

GECs typically work in the following fashion [1-4,12,13]. Initially, a population of CSs is randomly generated. Each CS is then evaluated and assigned a fitness based on a user-de‐ fined evaluation function. Next, parents are selected from the population based on their fitness and are allowed to produce offspring CSs. The offspring are then assigned a fitness and usually replace the worst performing CSs within the current population. This evolutionary process is then repeated until a user-specified stopping condition is satisfied. Figure 1 provides a flow‐

ods currently used within the biometrics research community.

well as directions for further research.

102 New Trends and Developments in Biometrics

chart of the typical GEC process.

**Figure 1.** Flowchart of a typical GEC.

**2.1. Genetic & Evolutionary Computation (GEC)**

**2. Background**

Within the field of biometrics, GEB applications (GEBAs) have been successfully applied to the areas of feature selection [25,30,31,33,35,36], feature weighting [25,30,31,37,38], feature extraction [26-29], and biometric-based security [9-11,39].

#### *2.2.1. GEBAs for Feature Selection and Weighting*

Concerning GEBAs for feature selection, Galbally et al. [35] developed Genetic Algorithms (GAs) [1] for feature selection and applied them in an effort to solve the signature verifica‐ tion problem. The GAs were applied to two training sets which were formed using the sig‐ natures of 330 subjects from the MCYT Signature database [40]. The first training set consisted of five signatures of each subject and the second training set consisted of 20 signa‐ tures of each subject. A test set was formed using the remaining signatures. Their results showed that the GAs, when compared to the baseline method that used all of the features, were able to reduce the number of features used while improving the recognition accuracy of the system.

Ramadan and Abdel-Kader [36] applied two GECs, Particle Swarm Optimization (PSO) [13] and a GA, for feature selection for a facial recognition problem. Images from the Cambridge ORL database [41] were used to form a training set, which consisted of four images of each of the 40 subjects, and a test set, which consisted of six images of each of the 40 subjects. Two baseline methods, the Discrete Cosine Transform (DCT) [42] and Discrete Wavelet Transform (DWT) [43] methods, were then used to extract the original set of features. Rama‐ dan and Abdel-Kader demonstrated that their GECs could increase the recognition accuracy over the baseline methods while using fewer features. In addition, the PSO used fewer fea‐ tures than the GA.

Kumar et al. [44] also applied two GECs, a Memetic Algorithm (MA) [45] and a GA, for feature selection for a face recognition system. The GECs were tested on two facial databases: the Cambridge ORL database [41] and a subset of 20 subjects from the YaleB database [46]. The following feature extraction techniques were used to create the original feature templates: Principal Component Analysis (PCA) [47], Linear Discriminant Analysis [48], and Kernel PCA [49]. The MA and GA were applied in an effort to reduce the dimensionality of the feature templates as well as to increase the recognition rate. Their results showed the GECs outper‐ formed the baseline methods, which used all of the extracted features, in terms of feature usage and recognition rate. Their results also showed that the MA outperformed the GA.

In [37], Abegaz et al. applied two GEBAs for feature selection and feature weighting for fa‐ cial recognition. The first GEBA, referred to as Genetic & Evolutionary Feature Selection (GEFeS), evolved subsets of biometric features also in an effort to increase the recognition rate while reducing the number of features used. The second GEBA, referred to as Genetic & Evolutionary Feature Weighting (GEFeW), evolved weights for the features. These GEBAs, which were instances of a Steady-State Genetic Algorithm (SSGA) [1-4] and Estimation of Distribution Algorithms (EDAs) [50], were applied to Eigenface[51] and LBP facial feature templates. Their results showed that GEFeS and GEFeW outperformed the baseline meth‐ ods, achieving higher recognition rates while using significantly fewer features. Their results also showed that the EDA instance of GEFeW was the best performing GEBA. In addition, the LBP instances outperformed the Eigenface instances.

Shelton et al. [26] then extended their work, by incorporating the machine learning techni‐ que known as cross validation [54], in an effort to evolve FEs that achieve high recognition rates, extract from a small region of the image, and generalize well to unseen subjects. The GEBA, known as GEFEML (GEFE-Machine Learning), was trained on a subset of images tak‐ en from the FRGC database. The evolved FEs were then tested on the images of unseen sub‐ jects taken from the FRGC database and Craniofacial Longitudinal Morphological Face (MORPH) database [53]. Their results showed that the GEFEML-evolved FEs, when com‐ pared to the SLBPM, used fewer patches, achieved comparable recognition rates on both da‐ tasets, and were significantly less expensive in terms of computational complexity. In addition, the resulting FEs generalized well to the unseen subjects from both databases.

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386 105

In [27], a two-stage process known as Darwinian Feature Extraction (DFE) was developed. This technique used FEs that were evolved by GEFEML to create Darwinian Feature Extrac‐ tors (dFEs). The first stage of DFE superimposes a set of FEs to create a hyper FE. This hyper FE is then used to create a probability distribution function (PDF) of the occurrence of ex‐ traction per pixel. In the second stage, the PDF is sampled to determine which pixels will be processed via the LBP method. The selected pixels are then grouped using a clustering proc‐ ess. Next, a number of cluster centers are randomly selected within an image, and the select‐ ed pixels are then assigned to their closest cluster centers. The SLBPM is then applied to the resulting clusters. Their results showed that dFEs, when compared to the FEs evolved by

GEFEML, achieved a higher recognition rate at a reduced computational complexity.

performed the other DFEC methods.

*2.2.3. Genetic & Evolutionary Biometric Security*

Adams et al. [28] extended the work of [27] by developing three types of Darwinian Feature Extraction – Clustering (DFEc). Unlike DFE, which uses random cluster centers, DFEC uses K-Means clustering [55] (DFEKM), Kohonen clustering [56] (DFEK) and a combination of the two (DFEK+KM and DFEKM+K) to create a user-specified number of clusters. These GEBAs were applied to datasets formed from subjects taken from the FRGC database. Their results showed that DFEKM and DFEK+KM performed as well as DFE in terms of accuracy, while DFEK and DFEKM+K performed worse than DFE. In addition, their results showed that DFEK+KM out‐

There has also been an increase in the use of GEBAs for biometric-based security, giving rise to a new field of study known as Genetic & Evolutionary Biometric Security (GEBS). Shelton et al. [9,10] introduced a number biometric-based access control protocols that used disposa‐ ble FEs and their associated feature vectors (FVs) to mitigate replay attacks on a facial recog‐ nition systems. In [9], the researchers showed that the FEs evolved by GEFEML and their associated FVs were unique from oneanother and achieved high recognition rate. As a re‐ sult, either the FEs or their associated FVs could be used to mitigate replay attacks by dis‐ posing of a particular FE or FV after usage. In [10], Shelton et al. extended their work by demonstrating that permuting the order of feature extraction resulted in the development of FVs with greater distinction from one another. The net effect of this being a further reduc‐ tion in the possibility of the occurrence of a successful biometric-based replay attack.In [11],

Alford et al. [38] compared the performances of GEFeS, GEFeW, and a hybrid feature selec‐ tion and weighting GEBA referred to as Genetic & Evolutionary Feature Weighting/Selec‐ tion (GEFeWS), for face-only, periocular-only, and face + periocular recognition. The GEBAs were implemented using a SSGA and an EDA, and the original feature sets were formed us‐ ing the Eigenface and LBP methods. Their results showed that the GEFeS, GEFeW, and GE‐ FeWS instances significantly outperformed the baseline methods in terms of recognition accuracy and feature usage. In addition, the EDA instance of GEFeWS outperformed the other GEBAs. Their results also showed that the performances of the GEBAs using multiple biometric modalities were better than those using only a single modality. In addition, the LBP instances performed better than the Eigenface instances.

Alford et al. [25] extended their work by developing a hybrid GEBA for feature selection and weighting that dramatically reduced the number of features necessary for recognition, increased the recognition rate, and also evolved feature masks (FMs) that generalized well to unseen subjects. The GEBA, known as GEFeWSML (GEFeWS – Machine Learning), was ap‐ plied to face-only, periocular-only, and face + periocular LBP templates formed for subjects within the Face Recognition Grand Challenge (FRGC) database [52]. Their results showed that GEFeWSML achieved higher recognition rates than the baseline methods, while using less than 50% of the extracted features. In addition, FMs evolved via the validation set per‐ formed better than those evolved via the training set.

In [31], Alford et al. evaluated the performance of GEFeWSML on a subset of images extract‐ ed from the Craniofacial Longitudinal Morphological Face (MORPH) [53] database. They al‐ so tested the cross-generalization ability of the resulting FMs by applying the FMs evolved for the FRGC datasets in [25] to the MORPH test set and applying the FMs evolved for the MORPH to the FRGC test set. Their results showed that the FMs evolved by GEFeWSML could also generalize well to unseen images from a different database.

#### *2.2.2. GEBAs for Feature Extraction*

Concerning GEBAs for feature extraction, Shelton et al. [29] developed a GEBA referred to as Genetic & Evolutionary Feature Extraction (GEFE), which evolved LBP-based feature extrac‐ tors (FEs) for facial recognition. Unlike the standard LBP method (SLBPM), GEFE evolved FEs that were allowed to overlap and only extracted from a subset of the image. Shelton et al. tested the performance of GEFE using two GECs, a SSGA and an EDA. They also evolved two types of FEs: (a) those that consisted of patches that were of non-uniform size and (b) those that consisted of patches that were of uniform size. Their results showed that the GEFE instances outperformed the SLBPM in terms of accuracy, feature usage, and computational complexity. In addition, the GEFE instances that evolved FEs composed of uniform sized patches outperformed the GEFE instances that evolved non-uniformed sized patches. Their results also showed that the EDA instances of GEFE outperformed the SSGA instances.

Shelton et al. [26] then extended their work, by incorporating the machine learning techni‐ que known as cross validation [54], in an effort to evolve FEs that achieve high recognition rates, extract from a small region of the image, and generalize well to unseen subjects. The GEBA, known as GEFEML (GEFE-Machine Learning), was trained on a subset of images tak‐ en from the FRGC database. The evolved FEs were then tested on the images of unseen sub‐ jects taken from the FRGC database and Craniofacial Longitudinal Morphological Face (MORPH) database [53]. Their results showed that the GEFEML-evolved FEs, when com‐ pared to the SLBPM, used fewer patches, achieved comparable recognition rates on both da‐ tasets, and were significantly less expensive in terms of computational complexity. In addition, the resulting FEs generalized well to the unseen subjects from both databases.

In [27], a two-stage process known as Darwinian Feature Extraction (DFE) was developed. This technique used FEs that were evolved by GEFEML to create Darwinian Feature Extrac‐ tors (dFEs). The first stage of DFE superimposes a set of FEs to create a hyper FE. This hyper FE is then used to create a probability distribution function (PDF) of the occurrence of ex‐ traction per pixel. In the second stage, the PDF is sampled to determine which pixels will be processed via the LBP method. The selected pixels are then grouped using a clustering proc‐ ess. Next, a number of cluster centers are randomly selected within an image, and the select‐ ed pixels are then assigned to their closest cluster centers. The SLBPM is then applied to the resulting clusters. Their results showed that dFEs, when compared to the FEs evolved by GEFEML, achieved a higher recognition rate at a reduced computational complexity.

Adams et al. [28] extended the work of [27] by developing three types of Darwinian Feature Extraction – Clustering (DFEc). Unlike DFE, which uses random cluster centers, DFEC uses K-Means clustering [55] (DFEKM), Kohonen clustering [56] (DFEK) and a combination of the two (DFEK+KM and DFEKM+K) to create a user-specified number of clusters. These GEBAs were applied to datasets formed from subjects taken from the FRGC database. Their results showed that DFEKM and DFEK+KM performed as well as DFE in terms of accuracy, while DFEK and DFEKM+K performed worse than DFE. In addition, their results showed that DFEK+KM out‐ performed the other DFEC methods.

#### *2.2.3. Genetic & Evolutionary Biometric Security*

ods, achieving higher recognition rates while using significantly fewer features. Their results also showed that the EDA instance of GEFeW was the best performing GEBA. In addition,

Alford et al. [38] compared the performances of GEFeS, GEFeW, and a hybrid feature selec‐ tion and weighting GEBA referred to as Genetic & Evolutionary Feature Weighting/Selec‐ tion (GEFeWS), for face-only, periocular-only, and face + periocular recognition. The GEBAs were implemented using a SSGA and an EDA, and the original feature sets were formed us‐ ing the Eigenface and LBP methods. Their results showed that the GEFeS, GEFeW, and GE‐ FeWS instances significantly outperformed the baseline methods in terms of recognition accuracy and feature usage. In addition, the EDA instance of GEFeWS outperformed the other GEBAs. Their results also showed that the performances of the GEBAs using multiple biometric modalities were better than those using only a single modality. In addition, the

Alford et al. [25] extended their work by developing a hybrid GEBA for feature selection and weighting that dramatically reduced the number of features necessary for recognition, increased the recognition rate, and also evolved feature masks (FMs) that generalized well to unseen subjects. The GEBA, known as GEFeWSML (GEFeWS – Machine Learning), was ap‐ plied to face-only, periocular-only, and face + periocular LBP templates formed for subjects within the Face Recognition Grand Challenge (FRGC) database [52]. Their results showed that GEFeWSML achieved higher recognition rates than the baseline methods, while using less than 50% of the extracted features. In addition, FMs evolved via the validation set per‐

In [31], Alford et al. evaluated the performance of GEFeWSML on a subset of images extract‐ ed from the Craniofacial Longitudinal Morphological Face (MORPH) [53] database. They al‐ so tested the cross-generalization ability of the resulting FMs by applying the FMs evolved for the FRGC datasets in [25] to the MORPH test set and applying the FMs evolved for the MORPH to the FRGC test set. Their results showed that the FMs evolved by GEFeWSML

Concerning GEBAs for feature extraction, Shelton et al. [29] developed a GEBA referred to as Genetic & Evolutionary Feature Extraction (GEFE), which evolved LBP-based feature extrac‐ tors (FEs) for facial recognition. Unlike the standard LBP method (SLBPM), GEFE evolved FEs that were allowed to overlap and only extracted from a subset of the image. Shelton et al. tested the performance of GEFE using two GECs, a SSGA and an EDA. They also evolved two types of FEs: (a) those that consisted of patches that were of non-uniform size and (b) those that consisted of patches that were of uniform size. Their results showed that the GEFE instances outperformed the SLBPM in terms of accuracy, feature usage, and computational complexity. In addition, the GEFE instances that evolved FEs composed of uniform sized patches outperformed the GEFE instances that evolved non-uniformed sized patches. Their results also showed that the EDA instances of GEFE outperformed the SSGA instances.

the LBP instances outperformed the Eigenface instances.

104 New Trends and Developments in Biometrics

LBP instances performed better than the Eigenface instances.

formed better than those evolved via the training set.

*2.2.2. GEBAs for Feature Extraction*

could also generalize well to unseen images from a different database.

There has also been an increase in the use of GEBAs for biometric-based security, giving rise to a new field of study known as Genetic & Evolutionary Biometric Security (GEBS). Shelton et al. [9,10] introduced a number biometric-based access control protocols that used disposa‐ ble FEs and their associated feature vectors (FVs) to mitigate replay attacks on a facial recog‐ nition systems. In [9], the researchers showed that the FEs evolved by GEFEML and their associated FVs were unique from oneanother and achieved high recognition rate. As a re‐ sult, either the FEs or their associated FVs could be used to mitigate replay attacks by dis‐ posing of a particular FE or FV after usage. In [10], Shelton et al. extended their work by demonstrating that permuting the order of feature extraction resulted in the development of FVs with greater distinction from one another. The net effect of this being a further reduc‐ tion in the possibility of the occurrence of a successful biometric-based replay attack.In [11], Shelton et al.showed that the dFEscreated by the DFE method could also be used to mitigate replay attacksas well.

[33]. Therefore, instead of having a histogram consisting of *2 <sup>P</sup>* bins, each histogram would now consist of only *P(P-1)+3* bins, where *P(P-1)* bins account for the uniform patterns with exactly two bit transitions, one bin represents the all zero bit pattern, another bin repre‐ sents the all ones bit pattern, and an additional bin is used to store the frequency of the non-

The histograms for each patch are then concatenated to form a feature template for the image, which consists of the number of bins, *P(P-1)+3*, times the number of patches used, *N*, features.

Genetic & Evolutionary Feature Weighting/Selection—Machine Learning (GEFeWSML) was developed by Alford et al. [25] for hybrid feature weighting and selection for facial and peri‐ ocular recognition. GEFeWSML evolves feature masks (FMs) that use a low percentage of fea‐ tures, achieve high recognition rates, and that generalize well to unseen instances. The technique, which is an instance of an Estimation of Distribution Algorithm (EDA) [50],

First, the set of available instances is split into a training set, a validation set, and a test set. The EDA then randomly generates an initial population of *Q* real-valued candidate FMs. Each

resents the set of mask values and where *fit <sup>i</sup>* represents the fitness value. The mask values are initially within the range [0..1]. Any mask value that is lower than 0.5 is set to 0 and the corresponding biometric feature would not be used during comparisons. Otherwise, the mask

Next, each candidate FM is applied to the probe and gallery templates within both the train‐ ing and validation sets. The resulting probe templates are then compared to each gallery template using the Weighted Manhattan Distance (wMD) formula shown in Equation 4,

tures in the biometric templates, *μ i,k* is a FM value, *k* represents the*k* th feature, and the func‐ tion *f WS* represents the process of feature weighting/selection as performed by GEFeWSML. The subject associated with the template within the gallery set with the smallest weighted

and *h <sup>l</sup>* are two templates that are being compared, *n* is the original number of fea‐

, *fiti* where *Mi* = *μi*,0, *μi*,1, ..., *μi*,*n*−1 rep‐

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386 107


**3.1. Genetic & Evolutionary Feature Weighting/Selection—Machine Learning**

, can be represented by a tuple *Mi*

Manhattan distance when compared to the probe is considered the match.

) =∑ *k*=0 *n*−1

, *hl* , *f mi*

value is used to weight the corresponding biometric feature.

*wM DWS* ( *hj*

uniform patterns.

**3. The GEBAs**

works as follows.

candidate FM, *fmi*

where *h <sup>j</sup>*

**(GEFeWSML)**

Adams et al. [39]developed a method for mitigating image reconstruction attacks from a compromised biometric gallery consisting of biometric templates extracted using the SLBPM. First, they demonstrated how a General Regression Neural Network [57] and a SSGA could be used to reconstruct facial images from the biometric templates. They then developed a neurogenetic method to distort the biometric templates, so that if reconstructed, they would not match with their original biometric templates.

#### **2.3. Local Binary Patterns Method (LBP)**

The Local Binary Patterns Method (LBP), which was developed by Ojala et al. [32], is a pop‐ ular texture descriptor that has been used in a variety of applications [25-28,30, 31, 34, 58-61]. Within the biometrics community, the LBP method has become a very popular fea‐ ture extraction technique [25-28, 30, 31, 58] due to its computational simplicity and its invari‐ ance against monotonic gray level changes [34].

The LBP works as follows. First, an image is segmented into a grid of evenly sized regions, which are referred to as patches. Within each patch, the LBP operator is then used to label each interior pixel by subtracting its intensity value,*i <sup>c</sup>*, from the intensity value of each of its *P* neighboring pixels at a radius *R*,*i <sup>p</sup>*, where *p* = *0,..., P-1*. A texture, which is essentially a bit pattern, is then formed, as shown in Equation 1, where if the difference between the intensi‐ ty values is negative, a zero is added to the bit string, otherwise a one.

$$T = \{ s(i\_0 - i\_c) \; , \; s(i\_1 - i\_c) \; , \; \dots \; , \; s(i\_{P-1} - i\_c) \} \tag{1}$$

$$\text{s.t.} \ (i\_p - i\_c) \ = \begin{cases} 0, \text{ if } i\_p < i\_c \\ 1, \text{ if } i\_p \ge i\_c \end{cases} \tag{2}$$

Each interior pixel can then be represented by a unique LBP code by assigning a binomial weight to the elements in the texture as follows:

$$LBP = \sum\_{p=0}^{P-1} s \text{ ( $i\_p - i\_c$ )} \text{ 2}^p \tag{3}$$

Associated with each patch is a histogram where each bin represents the frequency of a particular LBP code. Using a neighborhood size of *P*, there are *2 <sup>P</sup>* unique LBP codes. How‐ ever, Ojala et al. [32] showed that only a subset of the possible patterns, known as uniform patterns, are necessary to describe the texture of an image. For our research, uniform pat‐ terns are those texture patterns that have at most two one-to-zero or zero-to-one bit transi‐ tions when the texture is traversed circularly. These uniform patterns account for a high percentage of the resulting texture patterns and also contain the most texture information [33]. Therefore, instead of having a histogram consisting of *2 <sup>P</sup>* bins, each histogram would now consist of only *P(P-1)+3* bins, where *P(P-1)* bins account for the uniform patterns with exactly two bit transitions, one bin represents the all zero bit pattern, another bin repre‐ sents the all ones bit pattern, and an additional bin is used to store the frequency of the nonuniform patterns.

The histograms for each patch are then concatenated to form a feature template for the image, which consists of the number of bins, *P(P-1)+3*, times the number of patches used, *N*, features.

## **3. The GEBAs**

Shelton et al.showed that the dFEscreated by the DFE method could also be used to mitigate

Adams et al. [39]developed a method for mitigating image reconstruction attacks from a compromised biometric gallery consisting of biometric templates extracted using the SLBPM. First, they demonstrated how a General Regression Neural Network [57] and a SSGA could be used to reconstruct facial images from the biometric templates. They then developed a neurogenetic method to distort the biometric templates, so that if reconstructed,

The Local Binary Patterns Method (LBP), which was developed by Ojala et al. [32], is a pop‐ ular texture descriptor that has been used in a variety of applications [25-28,30, 31, 34, 58-61]. Within the biometrics community, the LBP method has become a very popular fea‐ ture extraction technique [25-28, 30, 31, 58] due to its computational simplicity and its invari‐

The LBP works as follows. First, an image is segmented into a grid of evenly sized regions, which are referred to as patches. Within each patch, the LBP operator is then used to label each interior pixel by subtracting its intensity value,*i <sup>c</sup>*, from the intensity value of each of its *P* neighboring pixels at a radius *R*,*i <sup>p</sup>*, where *p* = *0,..., P-1*. A texture, which is essentially a bit pattern, is then formed, as shown in Equation 1, where if the difference between the intensi‐

*<sup>c</sup>*), ..., *s*(*i*

Each interior pixel can then be represented by a unique LBP code by assigning a binomial

Associated with each patch is a histogram where each bin represents the frequency of a particular LBP code. Using a neighborhood size of *P*, there are *2 <sup>P</sup>* unique LBP codes. How‐ ever, Ojala et al. [32] showed that only a subset of the possible patterns, known as uniform patterns, are necessary to describe the texture of an image. For our research, uniform pat‐ terns are those texture patterns that have at most two one-to-zero or zero-to-one bit transi‐ tions when the texture is traversed circularly. These uniform patterns account for a high percentage of the resulting texture patterns and also contain the most texture information

*<sup>P</sup>*−<sup>1</sup> −*i*

<sup>ì</sup> <sup>&</sup>lt; - = <sup>í</sup> <sup>³</sup> <sup>î</sup> (2)

*<sup>c</sup>*) <sup>2</sup>*<sup>p</sup>* (3)

*p c*

*i i*

*p c*

*i i*

*<sup>c</sup>*)} (1)

they would not match with their original biometric templates.

ty values is negative, a zero is added to the bit string, otherwise a one.

<sup>0</sup> −*i <sup>c</sup>*), *s*(*i* <sup>1</sup> −*i*

*p c*

*LBP* =∑ *p*=0 *P*−1 *s* (*i <sup>p</sup>* −*i*

*si i*

0, if ( ) 1, if

*T* ={*s*(*i*

weight to the elements in the texture as follows:

**2.3. Local Binary Patterns Method (LBP)**

ance against monotonic gray level changes [34].

replay attacksas well.

106 New Trends and Developments in Biometrics

## **3.1. Genetic & Evolutionary Feature Weighting/Selection—Machine Learning (GEFeWSML)**

Genetic & Evolutionary Feature Weighting/Selection—Machine Learning (GEFeWSML) was developed by Alford et al. [25] for hybrid feature weighting and selection for facial and peri‐ ocular recognition. GEFeWSML evolves feature masks (FMs) that use a low percentage of fea‐ tures, achieve high recognition rates, and that generalize well to unseen instances. The technique, which is an instance of an Estimation of Distribution Algorithm (EDA) [50], works as follows.

First, the set of available instances is split into a training set, a validation set, and a test set. The EDA then randomly generates an initial population of *Q* real-valued candidate FMs. Each candidate FM, *fmi* , can be represented by a tuple *Mi* , *fiti* where *Mi* = *μi*,0, *μi*,1, ..., *μi*,*n*−1 rep‐ resents the set of mask values and where *fit <sup>i</sup>* represents the fitness value. The mask values are initially within the range [0..1]. Any mask value that is lower than 0.5 is set to 0 and the corresponding biometric feature would not be used during comparisons. Otherwise, the mask value is used to weight the corresponding biometric feature.

Next, each candidate FM is applied to the probe and gallery templates within both the train‐ ing and validation sets. The resulting probe templates are then compared to each gallery template using the Weighted Manhattan Distance (wMD) formula shown in Equation 4, where *h <sup>j</sup>* and *h <sup>l</sup>* are two templates that are being compared, *n* is the original number of fea‐ tures in the biometric templates, *μ i,k* is a FM value, *k* represents the*k* th feature, and the func‐ tion *f WS* represents the process of feature weighting/selection as performed by GEFeWSML. The subject associated with the template within the gallery set with the smallest weighted Manhattan distance when compared to the probe is considered the match.

$$\text{rawM } D\_{\text{WS}} \left( h\_{\restriction^{\nu}}, h\_{\restriction^{\nu}} \restriction f m\_i \right) = \sum\_{k=0}^{n-1} \restriction h\_{\restriction,k} - h\_{\restriction,k} \restriction f\_{\restriction \text{WS}} \left( u\_{i,k} \right) \tag{4}$$

$$f\_{WS}(u\_{\bar{i},k}) = \begin{cases} 0 & \text{, if } u\_{\bar{i},k} < 0.5\\ u\_{\bar{i},k} & \text{, if } u\_{\bar{i},k} \ge 0.5 \end{cases} \tag{5}$$

ated. The PPVs were then used to create a set of FEs. Each FE is created by selecting the top user-specified number ofpatches based on the patches PPV value, starting with all 24 patch‐ es as a baseline working down to the top performing patch. Ties were broken based on the patch position. A candidate FM is then created from each FE in the following manner. If a patch is selected to be present in the FE, then all of the features extracted by that patch will

Genetic &Evolutionary Feature Extraction—Machine Learning (GEFEML), which was devel‐ oped by Shelton et al. [26],evolves LBP-based FEs that result in high recognition accuracies, use a low percentage of features, and generalizes well to unseen images. Unlike the SLBPM, which uses non-overlapping uniform patches that extract from the entire image, GEFEML evolves FEs with overlapping uniform patches that only extract from a subset of an image.

As done with GEFeWSML, first the set of available instances is split into a training set, a vali‐ dation set, and a test set. GEFEML then randomly generates an initial population of *Q* candi‐

, can be represented as a six-tuple, *Xi*

first two elements in the tuple represent the *x* and *y* coordinates of the centers of the *N* possi‐ ble patches, where *Xi* = *xi*,0, *xi*,1, ..., *xi*,*<sup>N</sup>* <sup>−</sup>1 and *Yi* = *yi*,0, *yi*,1, ..., *yi*,*<sup>N</sup>* <sup>−</sup><sup>1</sup> . The next two ele‐ ments represent the widths and heights of the *N* patches respectively, where *Wi* = *wi*,0, *wi*,1, ..., *wi*,*<sup>N</sup>* <sup>−</sup><sup>1</sup> and*Hi* = *hi*,0, *hi*,1, ..., *hi*,*<sup>N</sup>* <sup>−</sup><sup>1</sup> . It is important to note that the size of the patches were uniform due to the results presented in [29], therefore, the weights and heights of each patch were equivalent. The final two elements in the six-tuple represent the masking values of each patch, where *Mi* = *mi*,0, *mi*,1, ..., *mi*,*<sup>N</sup>* <sup>−</sup>1 , and where*f <sup>i</sup>* is the fitness for

Next, the fitness function shown in Equation 7 is used to evaluate each candidate FE based on its performance on the training and validation sets, where γ represents the percentage of

training and validation set to create feature vectors (FVs) for the images. The resulting FVs for the probe and gallery sets are then compared using the Manhattan Distance measure. The subject associated with the template within the gallery set with the smallest Manhattan

The best performing FE on the validation set is then stored and will be referred to as *FE \**

with GEFeWSML, the top 50% best performing candidate FMs, based on their performance on the training set, are used to create a PDF. The PDF is then sampled to create *(1-α)Q* off‐ spring, where α = 5%. The offspring are then evaluated based on their performance on the

.

, *Yi* , *Wi* , *Hi* , *Mi*

*f <sup>i</sup>* =10*ε* + *γ* (7)

is applied to the probe and gallery sets in both the

, *f <sup>i</sup>* . The

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386 109

. As

be used for matching. Otherwise, no features will be extracted from the patch.

**3.3.Genetic & Evolutionary Feature Extraction-Machine Learning (GEFEML)**

The GEFEML technique, which is also an instance of an EDA [50], works as follows.

. The masking values determine whether a patch is activated or deactivated.

date FEs, where each FE, *fe <sup>i</sup>*

image space (measured in pixels) covered by *fe <sup>i</sup>*

distance when compared to the probe is considered the match.

To evaluate the candidate FEs, each*fe <sup>i</sup>*

*fe i*

The fitness function shown in Equation 6 is then used to evaluate each candidate FM based on its performance on the training and validation sets, where *ε* is the number of recognition errors that occurred when the candidate FM was applied to the probe and gallery templates, *m* is the number of features used by the candidate FM, and where *n* is the original number of features in the biometric templates.

$$
\hat{\varepsilon}\_i \hat{\varepsilon}\_i = 10\varepsilon + \frac{m}{n} \tag{6}
$$

The best performing candidate FM on the validation set is then stored and will be referred to as *FM\**. Next, the top 50% best performing candidate FMs, based on their performance on the training set, are used to create a probability density function (PDF). The PDF is then sampled to create (1-α)Q offspring, where α is the user-defined percentage (in this case, 25%) of FMs that are allowed to survive into the next generation (known as the elites). The offspring are then evaluated based on their performance on the training and validation sets. Any offspring whose performance on the validation set is better than that of FM\* will be‐ come the new FM\*. A new population is then formed using the (1-α)Q offspring and the αQ elites. This evolutionary process is then continued until a user-specified stopping condition has been satisfied. Once the condition has been met, the best performing FM from the popu‐ lation based on its performance on the training set, which will be referred to as FMts, as well as FM\* are returned. These FMs are then applied to the test set in order to determine how well they generalize to unseen instances.

#### **3.2.Reverse Engineered Feature Extraction (REFE)**

Reverse Engineered Feature Extraction (REFE) is a technique that creates feature extrac‐ tors (FEs) that obtain higher recognition rates than the SLBPM, while using a smaller num‐ ber of patches, as compared to the SLBPM. This technique works by first analyzing statistically the FMs evolved by GEFeWSML [25] to determine the percentage of features selected for use within each patch. For each FM, apatch percentage vector (PPV) is calculated by divid‐ ing the number of features used in a patch by the number of total features in a patch. For the results presented in this chapter, each image is segmented into 24 patches and each patch consisted of 59 bins. Therefore, the total number of features in a patch is 1416 (24 patches × 59 bins).

Each PPV is then evaluated at each patch position to create a corresponding FM. Because each image is segmented into 24 equally size patches, the PPV is evaluated based on the top 24 patches to the top patch, in terms of their percentage. In addition to the PPVs for the 30 best performing FMs, an average PPV (PPVavg) is also formed. PPVavg is the average percent‐ age of features used at each patch position for the 30 PPVs. Therefore, 31 total PPVs are cre‐ ated. The PPVs were then used to create a set of FEs. Each FE is created by selecting the top user-specified number ofpatches based on the patches PPV value, starting with all 24 patch‐ es as a baseline working down to the top performing patch. Ties were broken based on the patch position. A candidate FM is then created from each FE in the following manner. If a patch is selected to be present in the FE, then all of the features extracted by that patch will be used for matching. Otherwise, no features will be extracted from the patch.

## **3.3.Genetic & Evolutionary Feature Extraction-Machine Learning (GEFEML)**

,if 0.5 <sup>0</sup> , ( ) , ,if 0.5 , ,

<sup>ì</sup> <sup>&</sup>lt; <sup>ï</sup> <sup>=</sup> <sup>í</sup> <sup>³</sup> <sup>ï</sup>

The fitness function shown in Equation 6 is then used to evaluate each candidate FM based on its performance on the training and validation sets, where *ε* is the number of recognition errors that occurred when the candidate FM was applied to the probe and gallery templates, *m* is the number of features used by the candidate FM, and where *n* is the original number

*m*

The best performing candidate FM on the validation set is then stored and will be referred to as *FM\**. Next, the top 50% best performing candidate FMs, based on their performance on the training set, are used to create a probability density function (PDF). The PDF is then sampled to create (1-α)Q offspring, where α is the user-defined percentage (in this case, 25%) of FMs that are allowed to survive into the next generation (known as the elites). The offspring are then evaluated based on their performance on the training and validation sets. Any offspring whose performance on the validation set is better than that of FM\* will be‐ come the new FM\*. A new population is then formed using the (1-α)Q offspring and the αQ elites. This evolutionary process is then continued until a user-specified stopping condition has been satisfied. Once the condition has been met, the best performing FM from the popu‐ lation based on its performance on the training set, which will be referred to as FMts, as well as FM\* are returned. These FMs are then applied to the test set in order to determine how

Reverse Engineered Feature Extraction (REFE) is a technique that creates feature extrac‐ tors (FEs) that obtain higher recognition rates than the SLBPM, while using a smaller num‐ ber of patches, as compared to the SLBPM. This technique works by first analyzing statistically the FMs evolved by GEFeWSML [25] to determine the percentage of features selected for use within each patch. For each FM, apatch percentage vector (PPV) is calculated by divid‐ ing the number of features used in a patch by the number of total features in a patch. For the results presented in this chapter, each image is segmented into 24 patches and each patch consisted of 59 bins. Therefore, the total number of features in a patch is 1416 (24

Each PPV is then evaluated at each patch position to create a corresponding FM. Because each image is segmented into 24 equally size patches, the PPV is evaluated based on the top 24 patches to the top patch, in terms of their percentage. In addition to the PPVs for the 30 best performing FMs, an average PPV (PPVavg) is also formed. PPVavg is the average percent‐ age of features used at each patch position for the 30 PPVs. Therefore, 31 total PPVs are cre‐

*<sup>n</sup>* (6)

*i k f u WS i k <sup>u</sup> <sup>u</sup> i k i k*

î

*fiti* =10*ε* +

of features in the biometric templates.

108 New Trends and Developments in Biometrics

well they generalize to unseen instances.

patches × 59 bins).

**3.2.Reverse Engineered Feature Extraction (REFE)**

*u*

(5)

Genetic &Evolutionary Feature Extraction—Machine Learning (GEFEML), which was devel‐ oped by Shelton et al. [26],evolves LBP-based FEs that result in high recognition accuracies, use a low percentage of features, and generalizes well to unseen images. Unlike the SLBPM, which uses non-overlapping uniform patches that extract from the entire image, GEFEML evolves FEs with overlapping uniform patches that only extract from a subset of an image. The GEFEML technique, which is also an instance of an EDA [50], works as follows.

As done with GEFeWSML, first the set of available instances is split into a training set, a vali‐ dation set, and a test set. GEFEML then randomly generates an initial population of *Q* candi‐ date FEs, where each FE, *fe <sup>i</sup>* , can be represented as a six-tuple, *Xi* , *Yi* , *Wi* , *Hi* , *Mi* , *f <sup>i</sup>* . The first two elements in the tuple represent the *x* and *y* coordinates of the centers of the *N* possi‐ ble patches, where *Xi* = *xi*,0, *xi*,1, ..., *xi*,*<sup>N</sup>* <sup>−</sup>1 and *Yi* = *yi*,0, *yi*,1, ..., *yi*,*<sup>N</sup>* <sup>−</sup><sup>1</sup> . The next two ele‐ ments represent the widths and heights of the *N* patches respectively, where *Wi* = *wi*,0, *wi*,1, ..., *wi*,*<sup>N</sup>* <sup>−</sup><sup>1</sup> and*Hi* = *hi*,0, *hi*,1, ..., *hi*,*<sup>N</sup>* <sup>−</sup><sup>1</sup> . It is important to note that the size of the patches were uniform due to the results presented in [29], therefore, the weights and heights of each patch were equivalent. The final two elements in the six-tuple represent the masking values of each patch, where *Mi* = *mi*,0, *mi*,1, ..., *mi*,*<sup>N</sup>* <sup>−</sup>1 , and where*f <sup>i</sup>* is the fitness for *fe i* . The masking values determine whether a patch is activated or deactivated.

Next, the fitness function shown in Equation 7 is used to evaluate each candidate FE based on its performance on the training and validation sets, where γ represents the percentage of image space (measured in pixels) covered by *fe <sup>i</sup>* .

$$f\_i = 10\varepsilon + \gamma \tag{7}$$

To evaluate the candidate FEs, each*fe <sup>i</sup>* is applied to the probe and gallery sets in both the training and validation set to create feature vectors (FVs) for the images. The resulting FVs for the probe and gallery sets are then compared using the Manhattan Distance measure. The subject associated with the template within the gallery set with the smallest Manhattan distance when compared to the probe is considered the match.

The best performing FE on the validation set is then stored and will be referred to as *FE \** . As with GEFeWSML, the top 50% best performing candidate FMs, based on their performance on the training set, are used to create a PDF. The PDF is then sampled to create *(1-α)Q* off‐ spring, where α = 5%. The offspring are then evaluated based on their performance on the training and validation sets. Any offspring FE that outperforms *FE \** becomes the new *FE \** . Finally, a new population is formed using the *(1-α)Q*offspring and the *αQ* elites. This evolu‐ tionary process is then continued until a user-specified stopping condition has been satis‐ fied, after which two FEs are returned: the best performing FE on the training set, *FE ts*, and *FE \** . The returned FEs are then applied to the test set to evaluate how well they generalize to unseen instances in comparison to the SLBPM.

the Kohonen clustering process. Once each pixel has been assigned, the positions of the *K* centers are recalculated based on the average of the pixels assigned to that center. Once each of the centers has been repositioned, the distance between their old and new position is measured. When the positions of the centers remain constant, the K-Means clustering proc‐ ess is considered complete. The SLBPM is then applied to the clustered pixels and the histo‐ grams associated with each cluster are concatenated and are used as the feature vectors for

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386 111

For our experiments, images were selected from two diverse databases: the Face Recogni‐ tion Grand Challenge (FRGC) database [52] and the Craniofacial Longitudinal Morphologi‐ cal Face (MORPH) database [53]. The images within the two facial databases were acquired in different manners. The images within the FRGC database were collected in a controlled setting (i.e. controlled lighting, frontal pose, and neutral facial expression) and were ac‐ quired in a single sitting. In contrast, the images in the MORPH database were collected in an uncontrolled setting, were acquired over a period of time, and were of diverse ethnicities.

From the FRGC database, 489 subjects were selected. From the MORPH database, 300 sub‐ jects were selected. From the subjects selected from each database, three datasets were formed: a training set, a validation set, and a test set. The FRGC datasets were as follows: 163 of the subjects were used to form the training set, which will be referred to as FRGC-163trn; An additional 163 subjects were used to form the validation set, which will be referred to as FRGC-163val; The remaining 163 subjects were used to form the test set, which will be referred to as FRGC-163tst. The MORPH datasets were as follows: 100 subjects were used to form the training set, which will be referred to as MORPH-100trn; 100 subjects were used to form the validation set, which will be referred to as MORPH-100val; the remaining 100 subjects were used to form the test set, which will be referred to as MORPH-100tst.

For each of the facial datasets, three frontal facial images of each subject were selected and used to form the probe and gallery sets. The probe sets consisted of one image per subject, and the gallery sets consisted of the remaining two images per subject. The SLBPM was then used to extract 1416 (24 patches × 59 bins) facial features from each image and served as the

Using these datasets, we performed two experiments. For the first experiment, each GEBA was used to evolve FMs or FEs for the FRGC and MORPH facial templates within the train‐ ing sets. This will be referred to as '*Opt*' because we are attempting to maximize the recogni‐ tion accuracy while minimizing the percentage of features used. The resulting FMs/FEs were then applied to the test sets in order to evaluate how well they generalized to unseen sub‐ jects within the respective test sets. The application of the FM/FE that performed best on the training set (i.e. FMts and FEts) will be referred to as '*Opt-Gen*' because we are evaluating the

the given images.

**4.1. Databases**

**4. Experimental Results**

baseline for our experiments.

#### **3.4. Darwinian-Based Feature Extraction (DFE) using Random Clustering**

Darwinian-based Feature Extraction (DFE) [27] is a two-stage process for developing FEs to be used for facial recognition. In the first stage, a set of FEs evolved by GEFEML are superim‐ posed onto each other to create a hyper FE. From this hyper FE, a probability distribution function (PDF) is created which is essentially a two-dimensional array that represents the number of times a certain pixel is selected for use by the set of FEs. In the second stage of the process, a Darwinian-based FE (dFE) is created by sampling the PDF via k-tournament se‐ lection [62]. This dFE can be represented as a 3-tuple, *c*, *μ*, *ρ* , where c is the number of clusters in the dFE, µ is the selection pressure of tournament selection, and *ρ* is the resolu‐ tion of the clusters. To create a dFE, first a number of pixels,*c* ×*β* ×*ρ* , where β represents the user-specified number of pixels per cluster, are selected from the PDF for use in the cluster‐ ing process. Pixels are selected for use via k-tournament selection in which k = µ\*σ pixels, where σ represents the total number of positions in the PDF that have been processed at least once, compete for selection. The pixel with the greatest consistency (i.e. the pixel that was used the most by the set of FEs) will be considered the winner of the tournament and will not be allowed to win again. Next, *c* locations within an image are randomly selected to serve as cluster centers. The pixels that were selected for use are then assigned to their clos‐ est cluster center. The SLBPM is then applied to the clustered pixels in a manner similar to how the SLBPM is traditionally applied to pixels within a patch. Afterwards, the histograms associated with each cluster are concatenated and are used as the feature vectors for the giv‐ en images.

#### **3.5. Darwinian-Based Feature Extraction using Kohonen and K-Means Clustering (DFEK +KM)**

Darwinian Feature Extraction-Clustering [28] usingK-means[55] and Kohonen[56] clustering (DFEK+KM) works as follows. First, Kohonen clustering is applied in the following manner. The algorithm iterates though each of the pixels selected for use, moving the nearest cluster center towards the given pixel. The distance that the center is moved is based on a user specified learning rate. For example, given a learning rate of 0.25, the magnitude that the center is moved would be 25% of the distance between the current pixel and the center. Af‐ ter iterating through all of the pixels, the distance between each center's starting and ending position is calculated. If the locations of the centers have not moved or a user-specified num‐ ber of iterations have occurred, the Kohonen clustering process is halted. K-Means cluster‐ ing is then applied to the resulting clusters in the following manner.First, each pixel selected for use via tournament selection is assigned to one of the *K* clusters centers determined by the Kohonen clustering process. Once each pixel has been assigned, the positions of the *K* centers are recalculated based on the average of the pixels assigned to that center. Once each of the centers has been repositioned, the distance between their old and new position is measured. When the positions of the centers remain constant, the K-Means clustering proc‐ ess is considered complete. The SLBPM is then applied to the clustered pixels and the histo‐ grams associated with each cluster are concatenated and are used as the feature vectors for the given images.

## **4. Experimental Results**

#### **4.1. Databases**

.

training and validation sets. Any offspring FE that outperforms *FE \** becomes the new *FE \**

Finally, a new population is formed using the *(1-α)Q*offspring and the *αQ* elites. This evolu‐ tionary process is then continued until a user-specified stopping condition has been satis‐ fied, after which two FEs are returned: the best performing FE on the training set, *FE ts*, and

. The returned FEs are then applied to the test set to evaluate how well they generalize to

Darwinian-based Feature Extraction (DFE) [27] is a two-stage process for developing FEs to be used for facial recognition. In the first stage, a set of FEs evolved by GEFEML are superim‐ posed onto each other to create a hyper FE. From this hyper FE, a probability distribution function (PDF) is created which is essentially a two-dimensional array that represents the number of times a certain pixel is selected for use by the set of FEs. In the second stage of the process, a Darwinian-based FE (dFE) is created by sampling the PDF via k-tournament se‐ lection [62]. This dFE can be represented as a 3-tuple, *c*, *μ*, *ρ* , where c is the number of clusters in the dFE, µ is the selection pressure of tournament selection, and *ρ* is the resolu‐ tion of the clusters. To create a dFE, first a number of pixels,*c* ×*β* ×*ρ* , where β represents the user-specified number of pixels per cluster, are selected from the PDF for use in the cluster‐ ing process. Pixels are selected for use via k-tournament selection in which k = µ\*σ pixels, where σ represents the total number of positions in the PDF that have been processed at least once, compete for selection. The pixel with the greatest consistency (i.e. the pixel that was used the most by the set of FEs) will be considered the winner of the tournament and will not be allowed to win again. Next, *c* locations within an image are randomly selected to serve as cluster centers. The pixels that were selected for use are then assigned to their clos‐ est cluster center. The SLBPM is then applied to the clustered pixels in a manner similar to how the SLBPM is traditionally applied to pixels within a patch. Afterwards, the histograms associated with each cluster are concatenated and are used as the feature vectors for the giv‐

**3.5. Darwinian-Based Feature Extraction using Kohonen and K-Means Clustering (DFEK**

Darwinian Feature Extraction-Clustering [28] usingK-means[55] and Kohonen[56] clustering (DFEK+KM) works as follows. First, Kohonen clustering is applied in the following manner. The algorithm iterates though each of the pixels selected for use, moving the nearest cluster center towards the given pixel. The distance that the center is moved is based on a user specified learning rate. For example, given a learning rate of 0.25, the magnitude that the center is moved would be 25% of the distance between the current pixel and the center. Af‐ ter iterating through all of the pixels, the distance between each center's starting and ending position is calculated. If the locations of the centers have not moved or a user-specified num‐ ber of iterations have occurred, the Kohonen clustering process is halted. K-Means cluster‐ ing is then applied to the resulting clusters in the following manner.First, each pixel selected for use via tournament selection is assigned to one of the *K* clusters centers determined by

**3.4. Darwinian-Based Feature Extraction (DFE) using Random Clustering**

*FE \**

en images.

**+KM)**

unseen instances in comparison to the SLBPM.

110 New Trends and Developments in Biometrics

For our experiments, images were selected from two diverse databases: the Face Recogni‐ tion Grand Challenge (FRGC) database [52] and the Craniofacial Longitudinal Morphologi‐ cal Face (MORPH) database [53]. The images within the two facial databases were acquired in different manners. The images within the FRGC database were collected in a controlled setting (i.e. controlled lighting, frontal pose, and neutral facial expression) and were ac‐ quired in a single sitting. In contrast, the images in the MORPH database were collected in an uncontrolled setting, were acquired over a period of time, and were of diverse ethnicities.

From the FRGC database, 489 subjects were selected. From the MORPH database, 300 sub‐ jects were selected. From the subjects selected from each database, three datasets were formed: a training set, a validation set, and a test set. The FRGC datasets were as follows: 163 of the subjects were used to form the training set, which will be referred to as FRGC-163trn; An additional 163 subjects were used to form the validation set, which will be referred to as FRGC-163val; The remaining 163 subjects were used to form the test set, which will be referred to as FRGC-163tst. The MORPH datasets were as follows: 100 subjects were used to form the training set, which will be referred to as MORPH-100trn; 100 subjects were used to form the validation set, which will be referred to as MORPH-100val; the remaining 100 subjects were used to form the test set, which will be referred to as MORPH-100tst.

For each of the facial datasets, three frontal facial images of each subject were selected and used to form the probe and gallery sets. The probe sets consisted of one image per subject, and the gallery sets consisted of the remaining two images per subject. The SLBPM was then used to extract 1416 (24 patches × 59 bins) facial features from each image and served as the baseline for our experiments.

Using these datasets, we performed two experiments. For the first experiment, each GEBA was used to evolve FMs or FEs for the FRGC and MORPH facial templates within the train‐ ing sets. This will be referred to as '*Opt*' because we are attempting to maximize the recogni‐ tion accuracy while minimizing the percentage of features used. The resulting FMs/FEs were then applied to the test sets in order to evaluate how well they generalized to unseen sub‐ jects within the respective test sets. The application of the FM/FE that performed best on the training set (i.e. FMts and FEts) will be referred to as '*Opt-Gen*' because we are evaluating the generalization ability of these FMs/FEs. Similarly, the application of the best performing FM or FE on the validation set (i.e. FMts and FEts)will be referred to as '*Val-Gen*' because we are evaluating how well these FMs and FEs generalize.

To test the performance of DFEK+KM, the best hyper FEs from both FRGC (using *c =* 16) and MORPH (using *c=* 13) were used in the Kohonen clustering process. The results from Koho‐ nen clustering were then used for K-Means clustering. This process was performed for each of the 30 DFE runs. The resulting feature extractors were then applied to the FRGC and

The performances of these methods were separated into equivalence classes using ANOVA and t-tests based on their average recognition accuracy, the percentage of features used, and their computational complexity, which is the number of pixels processed or extracted by each method. Those methods that achieved higher recognition rates, used a lower percent‐ age of features, and had a reduced computational complexity, were considered to be best.

The results of our first experiment are shown in Tables 1 and 2. Table 1 shows the perform‐ ances of the GEBAs on the FRGC dataset. Table 2 shows the performances of the GEBAs on the MORPH dataset. Within each table, the first column represents the methods, where the asterisk denotes the performance of the GEBAs on the training set. The second column rep‐ resents the average recognition accuracy for each method. The third column represents the average computational complexity of each method, and the final column represents the average percentage of features used by each method.Note that for the SLBPM, the accuracy was deterministic. In addition, the results for REFE were the performances of the best per‐

With respect to GEFeWSML, the FMs evolved for the training set used an average of 45.02% of the features to achieve a higher recognition rate than the SLBPM. The FRGC Opt-Gen and Val-Gen performances show that the resulting FMs generalized well to the test set. When the generalization performances were compared, the Val-Gen performance was better in terms of accuracy, while the Opt-Gen performance used a lower percentage of features. This

REFE performed well on FRGC-163tst. The best performing FM created by analyzing the FMtss evolved by GEFeWSML achieved a 93.84% recognition rate while using only 29.16% (7

higher recognition rate of 94.46% using the same percentage of features. Although REFE did not outperform the SLBPM in terms of recognition accuracy, it's important to note that it significantly reduced the computational complexity of feature extraction by reducing the

In terms of recognition accuracy and feature usage, GEFEML outperformed the SLBPM on FRGC-163trn and FRGC-163tst. However, the average computational complexity of the FEs evolved by GEFEML was higher than the computational complexity of the SLBPM. This is due to the fact that the patches within the evolved FEs had a large amount of overlap. This overlap resulted in an increased number of pixels to be processed and, consequently, an in‐ crease in computational complexity. Though the computational complexities for FEs

s achieved a

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386 113

may be due to the need for more features for adequate generalization.

patches) of the features. The best performing FM created by analyzing the FM\*

MORPH test sets.

*4.2.1.Experiment I*

forming FMs.

number of patches from 24 to 7.

For the second experiment, we evaluated the cross-generalization ability of the resulting FMs/FEs for each GEBA. To do so, the FMs/FEs returned for the FRGC templates were ap‐ plied to the MORPH test set and the FMs/FEs returned for the MORPH templates were ap‐ plied to the FRGC test set.

#### **4.2. Results**

The results of our experiments were generated as follows. The SLBPM was applied to the training and test sets for the variousdatabases. Its performance served as the baseline for our experiments.

GEFeWSML was run 30 times on the FRGC, MORPH, and DLFW datasets. The EDA evolved a population of 20 FMs and always retained 5 (α=25%) elites. A maximum of 1000 function evaluations were allowed for each run, and at the end of each run, the best performing FM on the training set, *FM ts*, and the best performing FM on the validation set, *FM \** , were re‐ turned. These FMs were then applied to their respective test set in order to evaluate their generalization performances.

REFE analyzed the FMs evolved by GEFeWSMLfor the three datasets. The GEFeWSML evolved FMs were used to create a number of FEs. For each dataset, 62 PPVs were created (31 corre‐ sponding to the FMtss and 31 corresponding to the FM\* s). These PPVs were used to create FMs based on patch position and percentage of features used.The resulting FMs were then ap‐ plied to their respective test set in order to evaluate their generalization performances.

GEFEML was also run 30 times on the datasets, evolved a population size of 20 FEs, always retained one elite, and allowed a maximum of 1000 function evaluations per run. At the end of each run, GEFEML returned the best performing FE on the training set (*FE ts*), and the best performing FE with respect to the validation set (*FE\**). The resulting FEs were then applied to the test sets in order to evaluate how well they generalized to the test sets.

To construct a hyper FE, DFE used the 30 *FE \** s evolved by GEFEML for the FRGC, MORPH, and DLFW datasets. To test DFE on FRGC, dFEs were created with *c* values of 16, 12 and 8. These values were chosen based on the average number of patches activated in the 30 *FE \** s validated on FRGC-163val, which was 16; *c* values of 12 and 8 came from 75% and 50% of 16 clusters. For each *c*,a *ρ* of 100% to 10% was used, using every tenth percentage in between. A selection pressure from 100% to 0% was also used, using every tenth percentage.

To test DFE on MORPH, dFEs were created with *c* values of 13, 10, and 7. These values were chosen based on the average number of patches activated in the *30 FE \** s validated on MORPH-100val, which was 13; *c* values of 10 and 7 came from 75% and 50% of 16 clusters. For each *c*,a *ρ* of 100% to 10% was used, using every tenth percentage in between. A selec‐ tion pressure from 100% to 0% was also used, using every tenth percentage.

To test the performance of DFEK+KM, the best hyper FEs from both FRGC (using *c =* 16) and MORPH (using *c=* 13) were used in the Kohonen clustering process. The results from Koho‐ nen clustering were then used for K-Means clustering. This process was performed for each of the 30 DFE runs. The resulting feature extractors were then applied to the FRGC and MORPH test sets.

The performances of these methods were separated into equivalence classes using ANOVA and t-tests based on their average recognition accuracy, the percentage of features used, and their computational complexity, which is the number of pixels processed or extracted by each method. Those methods that achieved higher recognition rates, used a lower percent‐ age of features, and had a reduced computational complexity, were considered to be best.

## *4.2.1.Experiment I*

generalization ability of these FMs/FEs. Similarly, the application of the best performing FM or FE on the validation set (i.e. FMts and FEts)will be referred to as '*Val-Gen*' because we are

For the second experiment, we evaluated the cross-generalization ability of the resulting FMs/FEs for each GEBA. To do so, the FMs/FEs returned for the FRGC templates were ap‐ plied to the MORPH test set and the FMs/FEs returned for the MORPH templates were ap‐

The results of our experiments were generated as follows. The SLBPM was applied to the training and test sets for the variousdatabases. Its performance served as the baseline for our

GEFeWSML was run 30 times on the FRGC, MORPH, and DLFW datasets. The EDA evolved a population of 20 FMs and always retained 5 (α=25%) elites. A maximum of 1000 function evaluations were allowed for each run, and at the end of each run, the best performing FM

turned. These FMs were then applied to their respective test set in order to evaluate their

REFE analyzed the FMs evolved by GEFeWSMLfor the three datasets. The GEFeWSML evolved FMs were used to create a number of FEs. For each dataset, 62 PPVs were created (31 corre‐

based on patch position and percentage of features used.The resulting FMs were then ap‐ plied to their respective test set in order to evaluate their generalization performances.

GEFEML was also run 30 times on the datasets, evolved a population size of 20 FEs, always retained one elite, and allowed a maximum of 1000 function evaluations per run. At the end of each run, GEFEML returned the best performing FE on the training set (*FE ts*), and the best performing FE with respect to the validation set (*FE\**). The resulting FEs were then applied

and DLFW datasets. To test DFE on FRGC, dFEs were created with *c* values of 16, 12 and 8. These values were chosen based on the average number of patches activated in the 30 *FE \**

validated on FRGC-163val, which was 16; *c* values of 12 and 8 came from 75% and 50% of 16 clusters. For each *c*,a *ρ* of 100% to 10% was used, using every tenth percentage in between. A

To test DFE on MORPH, dFEs were created with *c* values of 13, 10, and 7. These values were

MORPH-100val, which was 13; *c* values of 10 and 7 came from 75% and 50% of 16 clusters. For each *c*,a *ρ* of 100% to 10% was used, using every tenth percentage in between. A selec‐

to the test sets in order to evaluate how well they generalized to the test sets.

selection pressure from 100% to 0% was also used, using every tenth percentage.

chosen based on the average number of patches activated in the *30 FE \**

tion pressure from 100% to 0% was also used, using every tenth percentage.

, were re‐

s

s validated on

s). These PPVs were used to create FMs

s evolved by GEFEML for the FRGC, MORPH,

on the training set, *FM ts*, and the best performing FM on the validation set, *FM \**

evaluating how well these FMs and FEs generalize.

plied to the FRGC test set.

112 New Trends and Developments in Biometrics

generalization performances.

sponding to the FMtss and 31 corresponding to the FM\*

To construct a hyper FE, DFE used the 30 *FE \**

**4.2. Results**

experiments.

The results of our first experiment are shown in Tables 1 and 2. Table 1 shows the perform‐ ances of the GEBAs on the FRGC dataset. Table 2 shows the performances of the GEBAs on the MORPH dataset. Within each table, the first column represents the methods, where the asterisk denotes the performance of the GEBAs on the training set. The second column rep‐ resents the average recognition accuracy for each method. The third column represents the average computational complexity of each method, and the final column represents the average percentage of features used by each method.Note that for the SLBPM, the accuracy was deterministic. In addition, the results for REFE were the performances of the best per‐ forming FMs.

With respect to GEFeWSML, the FMs evolved for the training set used an average of 45.02% of the features to achieve a higher recognition rate than the SLBPM. The FRGC Opt-Gen and Val-Gen performances show that the resulting FMs generalized well to the test set. When the generalization performances were compared, the Val-Gen performance was better in terms of accuracy, while the Opt-Gen performance used a lower percentage of features. This may be due to the need for more features for adequate generalization.

REFE performed well on FRGC-163tst. The best performing FM created by analyzing the FMtss evolved by GEFeWSML achieved a 93.84% recognition rate while using only 29.16% (7 patches) of the features. The best performing FM created by analyzing the FM\* s achieved a higher recognition rate of 94.46% using the same percentage of features. Although REFE did not outperform the SLBPM in terms of recognition accuracy, it's important to note that it significantly reduced the computational complexity of feature extraction by reducing the number of patches from 24 to 7.

In terms of recognition accuracy and feature usage, GEFEML outperformed the SLBPM on FRGC-163trn and FRGC-163tst. However, the average computational complexity of the FEs evolved by GEFEML was higher than the computational complexity of the SLBPM. This is due to the fact that the patches within the evolved FEs had a large amount of overlap. This overlap resulted in an increased number of pixels to be processed and, consequently, an in‐ crease in computational complexity. Though the computational complexities for FEs evolved by GEFEML were larger, they were selected due to their superior recognition accura‐ cy when compared to the SLBPM. When the performances of the FEtss were compared to the FE\* s in terms of recognition accuracy, the FE\* s (Val-Gen) performed statistically better than the FEtss (Opt-Gen).

while using only 41.66% of the features and significantly reducing the computational com‐ plexity by using only 10 patches. The best performing FM created with respect to the FM\*

also performed well on the test set, achieving a recognition accuracy that was slightly lower

GEFEML outperformed the SLBPM on MORPH-100trn and MORPH-100tst. The FEtss had an average 39.66% recognition accuracy when applied on MORPH-100trn, which was significant‐ ly better than the recognition accuracy achieved by the SLBPM. Like the FEs trained on FRGC, the average computational complexities of the FEs were also higher than for the SLBPM. The computational complexity of evolved FEs on the MORPH set were much higher due to the overlap as well as the large dimensions of the patches.However, FEs were still chosen primarily because of the recognition accuracy. In addition, when the

DFE outperformed the SLBPM in terms of recognition accuracy, computational complexity, and feature usage. DFE achieved a recognition rate of 59.73% while having a computational

DFEK+KM also outperformed the SLBPM in terms of recognition accuracy, computational complexity, and feature usage. This method achieved a recognition rate of 62.07% while

Comparing the generalization performances of the GEBAs, REFE outperformed the other GEBAs in terms of recognition accuracy, computational complexity, and feature usage.

**Method Accuracy Comp. Complexity % of Features**

having a computational complexity reduction of nearly 50% with respect to SLBPM.

SLBPM (Training) 27.00% 10379 100.00% SLBPM (Test) 70.00% 10379 100.00% GEFeWSML (Opt)\* 39.53% 10379 45.99% GEFeWSML (Opt-Gen) 61.67% 10379 45.99% GEFeWSML (Val-Gen) 64.30% 10379 48.61% REFE (Opt-Gen) 70.95% 4324 41.66% REFE (Val-Gen) 67.98% 3027 29.16% GEFEML (Opt)\* 39.66% 20034 58.33% GEFEML (Opt-Gen) 44.37% 20034 58.33% GEFEML (Val-Gen) 57.67% 11050 54.17% DFE (HFEvalon Test Set) 59.73% 5525 54.17% DFEK+KM (HFEvalon Test Set) 62.07% 5525 54.17%

than the SLBPM while using only 29.16% (7 patches) of the features.

performances of the FEtss were compared to the FE\*

complexity close to 50% less than the SLBPM.

FE\*

s performed better.

**Table 2.** MORPH Results.

s

115

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386

s in terms of recognition accuracy, the

For FRGC-163tst, DFE outperformed the SLBPM in terms of recognition accuracy as well as computational complexity. The recognition accuracy for DFE on FRGC-163tst was 97.24% and the computational complexity was approximately 10% less than the SLBPM.

DFEK+KM also outperformed SLBPM in terms of recognition accuracy and computational complexity. This method achieved a recognition accuracy of 97.20% on FRGC-163tst. Because the hyper FE from DFE was used, it had the same computational complexity as DFE.

When the performances of the GEBAs on the test set were compared in terms of accuracy, DFE and DFEK+KMoutperformed the other GEBAs. However, in terms of feature usage and computational complexity, REFE performed best.


**Table 1.** FRGC Results.

GEFeWSML outperformed the SLBPM on MORPH-100trn, using significantly fewer fea‐ tures while significantly increasing the recognition rate over the SLBPM. In addition, the resulting FMs generalized well to the test set. In terms of accuracy, the Val-Gen performances performed better statistically than the Opt-Gen performance. However, the Opt-Gen performances used a lower percentage of features in comparison to the Val-Gen performances.

The REFE created FMs performed well on MORPH-100tst. The best performing FM created with respect to the FMtss outperformed the SLBPM, achieving a 70.95% recognition rate while using only 41.66% of the features and significantly reducing the computational com‐ plexity by using only 10 patches. The best performing FM created with respect to the FM\* s also performed well on the test set, achieving a recognition accuracy that was slightly lower than the SLBPM while using only 29.16% (7 patches) of the features.

GEFEML outperformed the SLBPM on MORPH-100trn and MORPH-100tst. The FEtss had an average 39.66% recognition accuracy when applied on MORPH-100trn, which was significant‐ ly better than the recognition accuracy achieved by the SLBPM. Like the FEs trained on FRGC, the average computational complexities of the FEs were also higher than for the SLBPM. The computational complexity of evolved FEs on the MORPH set were much higher due to the overlap as well as the large dimensions of the patches.However, FEs were still chosen primarily because of the recognition accuracy. In addition, when the performances of the FEtss were compared to the FE\* s in terms of recognition accuracy, the FE\* s performed better.

DFE outperformed the SLBPM in terms of recognition accuracy, computational complexity, and feature usage. DFE achieved a recognition rate of 59.73% while having a computational complexity close to 50% less than the SLBPM.

DFEK+KM also outperformed the SLBPM in terms of recognition accuracy, computational complexity, and feature usage. This method achieved a recognition rate of 62.07% while having a computational complexity reduction of nearly 50% with respect to SLBPM.

Comparing the generalization performances of the GEBAs, REFE outperformed the other GEBAs in terms of recognition accuracy, computational complexity, and feature usage.


**Table 2.** MORPH Results.

evolved by GEFEML were larger, they were selected due to their superior recognition accura‐ cy when compared to the SLBPM. When the performances of the FEtss were compared to the

For FRGC-163tst, DFE outperformed the SLBPM in terms of recognition accuracy as well as computational complexity. The recognition accuracy for DFE on FRGC-163tst was 97.24%

DFEK+KM also outperformed SLBPM in terms of recognition accuracy and computational complexity. This method achieved a recognition accuracy of 97.20% on FRGC-163tst. Because

When the performances of the GEBAs on the test set were compared in terms of accuracy, DFE and DFEK+KMoutperformed the other GEBAs. However, in terms of feature usage and

**Method Accuracy Comp. Complexity % of Features**

and the computational complexity was approximately 10% less than the SLBPM.

the hyper FE from DFE was used, it had the same computational complexity as DFE.

SLBPM (Training) 92.02% 10379 100.00% SLBPM (Test) 95.09% 10379 100.00% GEFeWSML (Opt)\* 97.63% 10379 45.02% GEFeWSML(Opt-Gen) 94.42% 10379 45.02% GEFeWSML (Val-Gen) 94.58% 10379 48.45% REFE (Opt-Gen) 93.84% 3027 29.16% REFE (Val-Gen) 94.46% 3027 29.16% GEFEML (Opt)\* 96.85% 11776 66.67% GEFEML (Opt-Gen) 96.64% 11776 66.67% GEFEML (Val-Gen) 96.87% 11776 66.67% DFE (HFEvalon Test Set) 97.24% 9011 66.67% DFEK+KM 97.20% 9011 66.67%

GEFeWSML outperformed the SLBPM on MORPH-100trn, using significantly fewer fea‐ tures while significantly increasing the recognition rate over the SLBPM. In addition, the resulting FMs generalized well to the test set. In terms of accuracy, the Val-Gen performances performed better statistically than the Opt-Gen performance. However, the Opt-Gen performances used a lower percentage of features in comparison to the Val-

The REFE created FMs performed well on MORPH-100tst. The best performing FM created with respect to the FMtss outperformed the SLBPM, achieving a 70.95% recognition rate

s (Val-Gen) performed statistically better than

FE\*

the FEtss (Opt-Gen).

114 New Trends and Developments in Biometrics

**Table 1.** FRGC Results.

Gen performances.

s in terms of recognition accuracy, the FE\*

computational complexity, REFE performed best.

## *4.2.2. Experiment II*

Tables 3 and 4 show the results of our second experiment. Table 3 presents the cross-gener‐ alization performances of the MORPH FMs/FEs on the FRGC test set, while Table 4 presents the cross-generalization performances of the FRGC FMs/FEs on the MORPH test set. The first column of each table represents the methods, the second column represents the average recognition accuracy, the third column represents the average computational complexity of each method, and the final column represents the average percentage of features used by each method. As for Experiment I, for the SLBPM, the accuracy was deterministic and the results shown for REFE were the performances of the best performing FMs.

With respect to GEFEML, the recognition accuracyof the SLBPM was superior to the average

GEFEMLwas 50% less than the SLBPM. The Val-Gen recognition accuracy was also statistical‐ ly greater than the Opt-Gen accuracy, and the computational complexity was less for Val-

For DFE, the recognition accuracy was slightly lower than the accuracy of the SLBPM. How‐ ever, the percentage of features used by DFE, as well the computational complexity of DFE,

DFEK+KM performed well on FRGC-163tst, achieving a recognition rate slightly lower than the SLBPM while having a nearly 50% lower computational complexity and fea‐

Comparing the cross-generalization performances of the GEBAs, DFE cross-generalized best to the FRGC dataset. The dFEs achieved recognition rates comparable to the SLBPM, while

When the FMs evolved by GEFeWSMLfor FRGC were applied to the MORPH test set, they did not outperform the SLBPM, which was applied directly to the test set. However, the per‐ formances of the FMtss evolved for FRGC were better statistically than the Opt-Gen results presented in Table 2. In addition, the Val-Gen performances for both experiments were stat‐

The best performing FMs created by REFE for FRGC generalized well to the MORPH test set. The FMs outperformed the SLBPM in terms of accuracy, computational complexity, and

With respect to GEFEML, the recognition accuracyof the SLBPM was superior to the average

features used by GEFEML. The Val-Gen recognition accuracy was still statistically greater than the Opt-Gen accuracy, though the computational complexity remained the same. The recognition accuracy resulting from cross-generalizing FEs on the MORPH dataset is greater

The recognition accuracy of DFE was less than the recognition accuracy of the SLBPM. How‐ ever, the percentage of features used by DFE, as well as the computational complexity, was

DFEK+KM outperformed the SLBPM in terms of accuracy, computational complexity, and fea‐ ture usage. The GEBA achieved a 71.12% recognition rate, while using only 66.67% of the

Comparing the performances of the SLBPM and the GEBAs, REFE was the best performing technique, generalizing well to the test set, while processing less than 50% of the available

s. However, there is still a reduction in the percentage of

using significantly fewer features and requiring less processing time.

s. However, the percentage of features for

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386 117

accuracy of both the FEtss as well as the FE\*

Gen than for Opt-Gen.

ture usage.

istically equivalent.

accuracy of both the FEtssand FE\*

features and processing fewer pixels.

pixels and using less than 30% of the features.

than the accuracy of generalizing FEs in Experiment I.

far less than the percentage of feature for the SLBPM.

feature usage.

was far less than for the SLBPM.

With respect to GEFeWSML, the FMtss and FM\* s evolved for the MORPH dataset generalized well to the FRGC test set. Although the average accuracy waslower than the SLBPM and lower than the results presented in Table 1, it is important to note that the FMs were opti‐ mized for the MORPH dataset, while the SLBPM and FMs used for Experiment I were ap‐ plied directly to and trained on the FRGC test set. In addition, the FMs used less than 50% of the features in contrast to the SLBPM which used 100% of the features. When the Opt-Gen and Val-Gen performances were compared, in terms of accuracy, Val-Gen performed than Opt-Gen. However, Opt-Gen used fewer features.


**Table 3.** MORPH to FRGC Cross-Generalization Results.

The best performing FMs created by REFE for the MORPH dataset did not generalize well to FRGC-163tst. Neither FMs outperformed the SLBPM in terms of recognition accuracy. How‐ ever, the best performing FM with respect to the FMtss achieved a 72.97% recognition rate while using only 41.66% of the features and while having a more than 50% lower computa‐ tional complexity. The best performing FM with respect to the FM\* s achieved a 68.97% rec‐ ognition rate while using only 29.16% of the features and also having a 50% lower computational complexity. This reduction in features and processing time (in terms of com‐ putational complexity) is important to highlight, especially since the SLBPM was applied di‐ rectly to the test set.

With respect to GEFEML, the recognition accuracyof the SLBPM was superior to the average accuracy of both the FEtss as well as the FE\* s. However, the percentage of features for GEFEMLwas 50% less than the SLBPM. The Val-Gen recognition accuracy was also statistical‐ ly greater than the Opt-Gen accuracy, and the computational complexity was less for Val-Gen than for Opt-Gen.

*4.2.2. Experiment II*

116 New Trends and Developments in Biometrics

Tables 3 and 4 show the results of our second experiment. Table 3 presents the cross-gener‐ alization performances of the MORPH FMs/FEs on the FRGC test set, while Table 4 presents the cross-generalization performances of the FRGC FMs/FEs on the MORPH test set. The first column of each table represents the methods, the second column represents the average recognition accuracy, the third column represents the average computational complexity of each method, and the final column represents the average percentage of features used by each method. As for Experiment I, for the SLBPM, the accuracy was deterministic and the

well to the FRGC test set. Although the average accuracy waslower than the SLBPM and lower than the results presented in Table 1, it is important to note that the FMs were opti‐ mized for the MORPH dataset, while the SLBPM and FMs used for Experiment I were ap‐ plied directly to and trained on the FRGC test set. In addition, the FMs used less than 50% of the features in contrast to the SLBPM which used 100% of the features. When the Opt-Gen and Val-Gen performances were compared, in terms of accuracy, Val-Gen performed than

**Method Accuracy Comp. Complexity % of Features**

SLBPM (Test) 95.09% 10379 100.00% GEFeWSML(Opt-Gen) 92.70% 10379 45.99% GEFeWSML (Val-Gen) 93.72% 10379 48.61% REFE (Opt-Gen) 72.97% 4324 41.66% REFE (Val-Gen) 68.97% 3027 29.16% GEFEML (Opt-Gen) 88.41% 20034 58.33% GEFEML (Val-Gen) 93.37% 11050 54.17% DFE (HFEvalon Test set) 94.66% 5525 54.17% DFEK+KM 93.82% 5525 54.17%

The best performing FMs created by REFE for the MORPH dataset did not generalize well to FRGC-163tst. Neither FMs outperformed the SLBPM in terms of recognition accuracy. How‐ ever, the best performing FM with respect to the FMtss achieved a 72.97% recognition rate while using only 41.66% of the features and while having a more than 50% lower computa‐

ognition rate while using only 29.16% of the features and also having a 50% lower computational complexity. This reduction in features and processing time (in terms of com‐ putational complexity) is important to highlight, especially since the SLBPM was applied di‐

tional complexity. The best performing FM with respect to the FM\*

s evolved for the MORPH dataset generalized

s achieved a 68.97% rec‐

results shown for REFE were the performances of the best performing FMs.

With respect to GEFeWSML, the FMtss and FM\*

Opt-Gen. However, Opt-Gen used fewer features.

**Table 3.** MORPH to FRGC Cross-Generalization Results.

rectly to the test set.

For DFE, the recognition accuracy was slightly lower than the accuracy of the SLBPM. How‐ ever, the percentage of features used by DFE, as well the computational complexity of DFE, was far less than for the SLBPM.

DFEK+KM performed well on FRGC-163tst, achieving a recognition rate slightly lower than the SLBPM while having a nearly 50% lower computational complexity and fea‐ ture usage.

Comparing the cross-generalization performances of the GEBAs, DFE cross-generalized best to the FRGC dataset. The dFEs achieved recognition rates comparable to the SLBPM, while using significantly fewer features and requiring less processing time.

When the FMs evolved by GEFeWSMLfor FRGC were applied to the MORPH test set, they did not outperform the SLBPM, which was applied directly to the test set. However, the per‐ formances of the FMtss evolved for FRGC were better statistically than the Opt-Gen results presented in Table 2. In addition, the Val-Gen performances for both experiments were stat‐ istically equivalent.

The best performing FMs created by REFE for FRGC generalized well to the MORPH test set. The FMs outperformed the SLBPM in terms of accuracy, computational complexity, and feature usage.

With respect to GEFEML, the recognition accuracyof the SLBPM was superior to the average accuracy of both the FEtssand FE\* s. However, there is still a reduction in the percentage of features used by GEFEML. The Val-Gen recognition accuracy was still statistically greater than the Opt-Gen accuracy, though the computational complexity remained the same. The recognition accuracy resulting from cross-generalizing FEs on the MORPH dataset is greater than the accuracy of generalizing FEs in Experiment I.

The recognition accuracy of DFE was less than the recognition accuracy of the SLBPM. How‐ ever, the percentage of features used by DFE, as well as the computational complexity, was far less than the percentage of feature for the SLBPM.

DFEK+KM outperformed the SLBPM in terms of accuracy, computational complexity, and fea‐ ture usage. The GEBA achieved a 71.12% recognition rate, while using only 66.67% of the features and processing fewer pixels.

Comparing the performances of the SLBPM and the GEBAs, REFE was the best performing technique, generalizing well to the test set, while processing less than 50% of the available pixels and using less than 30% of the features.


Figures 2 and 3 show the ROC curves of the SLBPM, and the best performing FMts, FM\*

that the FEs perform better than the SLBPM and the FMs on both datasets.

**Figure 2.** ROC curves for the SLBPM and the best performing FMs and FEs on FRGC-163tst.

**Figure 3.** ROC curves for the SLBPM and the best performing FMs and FEs on MORPH-163tst.

perform comparable to the SLBPM while using significantly fewer features. It also appears

In addition, Figure 4 shows the graph of the accuracy of the best performing REFE created FMs on FRGC-163tst and MORPH-100tst. By using only the patches that correspond to the highest feature usages, one can achieve recognition accuracies better than using all 24 patch‐

on FRGC-163tst and MORPH-100tst respectively. The evolved FMs and FEs seem to

and FE\*

es as done by the SLBPM.

, FEts,

119

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386

**Table 4.** FRGC to MORPH Cross-Generalization Results.

#### **5. Discussion**

In Experiment I, we showed that the GEBAs could evolve FEs and FMs that achieve accura‐ cies comparable to the performance of the SLBPM on the test sets, while using significantly fewer features and a using a lower computational complexity. To further analyze the per‐ formances of the GEBAs, the Receiver Operator Characteristic (ROC) curve for the perform‐ ance of the best FMs/FEs on the test sets were plotted in Figures 2 and 3. The ROC curves for GEFeWSML were created using the normalized Weighted Manhattan Distance(NwMD) for‐ mula shown in Equation 8, and the ROC curves for GEFEML were created using the normal‐ ized Manhattan Distance(NMD) formula shown in Equation 9. For these formulas, *h <sup>i</sup>* and *h j* are the feature templates being compared, *fm <sup>l</sup>* is the GEFeWSMLevolved FM, *z* is the *z th* fea‐ turewithin the templates, *n* is the total number of extracted features, andthe function*f WS ,* asshown in Equation 5, represents the process of feature weighting/selection as performed by GEFeWSML. The ROC charts plot the true accept rate (y-axis) to the false accept rate (xaxis). For a given threshold from 0.0 to 1.0, all images in the probe dataset were compared to all images in the gallery dataset. If the NMD of the probe image and gallery image was be‐ low the threshold, it was considered a match. The true accept rate was increased if the two matching images were of the same subject, while the false accept rate was increased if the images were not of the same subject.

$$\text{NwM } \mathbf{D}\_{\text{WS}}(\mathbf{h}\_{i'}, \mathbf{h}\_{j'}, f \, \mathbf{m}\_l) = \sum\_{z=0}^{n-1} \frac{\|\mathbf{h}\_{i,z} - \mathbf{h}\_{j,z}\| \, f \, \mathbf{y}\_{\text{WS}}(\mathbf{u}\_{i,z})}{\max(\mathbf{h}\_{i,z'}, \mathbf{h}\_{j,z}) f \, \mathbf{y}\_{\text{WS}}(\mathbf{u}\_{i,z})} \tag{8}$$

$$\text{NMD}(h\_{j,\nu} \ h\_j) = \sum\_{z=0}^{n-1} \frac{|\ h\_{j,z} - h\_{\ j,z}|}{\max(h\_{j,z} - h\_{\ j,z})} \tag{9}$$

Figures 2 and 3 show the ROC curves of the SLBPM, and the best performing FMts, FM\* , FEts, and FE\* on FRGC-163tst and MORPH-100tst respectively. The evolved FMs and FEs seem to perform comparable to the SLBPM while using significantly fewer features. It also appears that the FEs perform better than the SLBPM and the FMs on both datasets.

**Method Accuracy Comp. Complexity % of Features**

SLBPM (Test) 70.00% 10379 100.00% GEFeWSML (Opt-Gen) 65.19% 10379 45.02% GEFeWSML (Val-Gen) 65.60% 10379 48.45% REFE (Opt-Gen) 72.98% 3027 29.16% REFE (Val-Gen) 71.14% 3027 29.16% GEFEML (Opt-Gen) 64.76% 11776 66.67% GEFEML (Val-Gen) 65.63% 11776 66.67% DFE (HFEvalon Test set) 62.57% 9011 66.67% DFEK+KM 71.12% 9011 66.67%

In Experiment I, we showed that the GEBAs could evolve FEs and FMs that achieve accura‐ cies comparable to the performance of the SLBPM on the test sets, while using significantly fewer features and a using a lower computational complexity. To further analyze the per‐ formances of the GEBAs, the Receiver Operator Characteristic (ROC) curve for the perform‐ ance of the best FMs/FEs on the test sets were plotted in Figures 2 and 3. The ROC curves for GEFeWSML were created using the normalized Weighted Manhattan Distance(NwMD) for‐ mula shown in Equation 8, and the ROC curves for GEFEML were created using the normal‐ ized Manhattan Distance(NMD) formula shown in Equation 9. For these formulas, *h <sup>i</sup>*

turewithin the templates, *n* is the total number of extracted features, andthe function*f WS ,* asshown in Equation 5, represents the process of feature weighting/selection as performed by GEFeWSML. The ROC charts plot the true accept rate (y-axis) to the false accept rate (xaxis). For a given threshold from 0.0 to 1.0, all images in the probe dataset were compared to all images in the gallery dataset. If the NMD of the probe image and gallery image was be‐ low the threshold, it was considered a match. The true accept rate was increased if the two matching images were of the same subject, while the false accept rate was increased if the

and *h*

is the GEFeWSMLevolved FM, *z* is the *z th* fea‐

max(*hi*,*z*, *<sup>h</sup> <sup>j</sup>*,*z*) *<sup>f</sup> WS* (*ui*,*z*) (8)

max(*hi*,*<sup>z</sup>* <sup>−</sup>*<sup>h</sup> <sup>j</sup>*,*z*) (9)

**Table 4.** FRGC to MORPH Cross-Generalization Results.

118 New Trends and Developments in Biometrics

are the feature templates being compared, *fm <sup>l</sup>*

*NwM DWS* (*hi*

, *hj* , *f ml*

*NMD*(*hi*

, *hj* )=∑ *z*=0

)=∑ *z*=0

*<sup>n</sup>*−<sup>1</sup> |*hi*,*<sup>z</sup>* −*h <sup>j</sup>*,*<sup>z</sup>* | *f WS* (*ui*,*z*)

*<sup>n</sup>*−<sup>1</sup> |*hi*,*<sup>z</sup>* −*h <sup>j</sup>*,*<sup>z</sup>* |

images were not of the same subject.

**5. Discussion**

*j*

In addition, Figure 4 shows the graph of the accuracy of the best performing REFE created FMs on FRGC-163tst and MORPH-100tst. By using only the patches that correspond to the highest feature usages, one can achieve recognition accuracies better than using all 24 patch‐ es as done by the SLBPM.

**Figure 2.** ROC curves for the SLBPM and the best performing FMs and FEs on FRGC-163tst.

**Figure 3.** ROC curves for the SLBPM and the best performing FMs and FEs on MORPH-163tst.

**Author details**

Aniesha Alford1

Jonathan Turner2

John Kelly1

ty, USA

**References**

York.

& Sons, Ltd.

Intelligence,. IEEE Press.

Springer-Verlag , 3-23.

Agents. MIT Press Cambridge, MA.

Verlag 978-3-54048-582-7 pages., 628.

New York, Inc., Secaucus, NJ, USA.

*nual Conference on Systems Engineering Research (CSER)*.

, Joseph Shelton2\*, Joshua Adams2

, Robert Benson2

1 Department of Electrical and Computer Engineering, North Carolina A&T State Universi‐

[1] Davis, L. (1991). Handbook of genetic algorithms. Van Nostrand Reinhold, New

[2] Dozier, G., Homaifar, A., Tunstel, E., & Battle, D. (2001). An Introduction to Evolu‐ tionary Computation (Chapter 17). Intelligent Control Systems Using Soft Comput‐

[3] Engelbrecht, A. P. (2007). Computational Intelligence: An Introduction,. John Wiley

[4] Fogel, D. (1995). Evolutionary Computation: Toward a New Philosophy of Machine

[5] Nolfi, S., & Floreano, D. (2000). Evolutionary robotics- the biology, intelligence, and technology of self-organizing machines. In: Intelligent Robotics and Automation

[6] Dahal, K. P., Tan, K. C., & Cowling, P. I. (2007). Evolutionary Scheduling. Springer

[7] Miranda, E. R., & Biles, J. A. (2007). Evolutionary Computer Music. Springer-Verlag,

[8] Dasgupta, D. (1999). An overview of artificial immune systems and their applica‐ tions. In Dasgupta, D., editor, Artificial Immune System and Their Applications

[9] Shelton, J., Bryant, K., Abrams, S., Small, L., Adams, J., Leflore, D., Alford, A., Rica‐ nek, K., & Dozier, G. (2012). Genetic & Evolutionary Biometric Security: Disposable Feature Extractors for Mitigating Biometric Replay Attacks. *Proceedings of the 10th An‐*

ing Methodologies, A. Zilouchian& M. Jamshidi (Eds.), CRC press , 365-380.

2 Department of Computer Science, North Carolina A&T State University, USA

, Vincent McLean2

\*Address all correspondence to: jashelt1@ncat.edu

, Derrick LeFlore2

, Gerry Dozier2

, Michael Payne2

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386

, Kelvin Bryant2

,

and

121

**Figure 4.** Performance of the best REFE feature masks for FRGC and MORPH test sets.

## **6. Conclusions and Further Work**

In conclusion, we presented several GEBAs for feature selection/weighting and for feature extraction. The GEBAs were applied to images of subjects taken from two facial. The result‐ ing FMs/FEs developed for each database were then applied to instances of each unrelated database to test the notion of cross-generalization. Our results showed that the GEBAs ach‐ ieved recognition rates comparable to the SLBPM while using significantly fewer features and a considerably lower computational complexity.

Further work will be devoted towards the use of these GEBAs to develop more secure bio‐ metric systems.

## **Acknowledgements**

This research was funded by the Office of the Director of National Intelligence (ODNI), Cen‐ ter for Academic Excellence (CAE) for the multi-university Center for Advanced Studies in Identity Sciences (CASIS) and by the National Science Foundation (NSF) Science & Technol‐ ogy Center: Bio/computational Evolution in Action CONsortium (BEACON). The authors would like to thank the ODNI and the NSF for their support of this research.

## **Author details**

Aniesha Alford1 , Joseph Shelton2\*, Joshua Adams2 , Derrick LeFlore2 , Michael Payne2 , Jonathan Turner2 , Vincent McLean2 , Robert Benson2 , Gerry Dozier2 , Kelvin Bryant2 and John Kelly1

\*Address all correspondence to: jashelt1@ncat.edu

1 Department of Electrical and Computer Engineering, North Carolina A&T State Universi‐ ty, USA

2 Department of Computer Science, North Carolina A&T State University, USA

## **References**

**Figure 4.** Performance of the best REFE feature masks for FRGC and MORPH test sets.

In conclusion, we presented several GEBAs for feature selection/weighting and for feature extraction. The GEBAs were applied to images of subjects taken from two facial. The result‐ ing FMs/FEs developed for each database were then applied to instances of each unrelated database to test the notion of cross-generalization. Our results showed that the GEBAs ach‐ ieved recognition rates comparable to the SLBPM while using significantly fewer features

Further work will be devoted towards the use of these GEBAs to develop more secure bio‐

This research was funded by the Office of the Director of National Intelligence (ODNI), Cen‐ ter for Academic Excellence (CAE) for the multi-university Center for Advanced Studies in Identity Sciences (CASIS) and by the National Science Foundation (NSF) Science & Technol‐ ogy Center: Bio/computational Evolution in Action CONsortium (BEACON). The authors

would like to thank the ODNI and the NSF for their support of this research.

**6. Conclusions and Further Work**

120 New Trends and Developments in Biometrics

metric systems.

**Acknowledgements**

and a considerably lower computational complexity.


[10] Shelton, J., Dozier, G., Adams, J., & Alford, A. (2012). Permutation-Based Biometric Authentication Protocols for Mitigating Replay Attacks. Paper presented at Proceed‐ ings of the 2012 IEEE World Congress on Computational Intelligence (WCCI).

[24] Dahal, K. P., Tan, K. C., & Cowling, P. I. (2007). Evolutionary Scheduling. *Springer*

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386 123

[25] Alford, A., Steed, C., Jeffrey, M., Sweet, D., Shelton, J., Small, L., Leflore, D., Dozier, G., Bryant, K., Abegaz, T., Kelly, J. C., & Ricanek, K. (2012). Genetic & Evolutionary Biometrics: Hybrid Feature Selection and Weighting for a Multi-Modal Biometric

[26] Shelton, J., Alford, A., Abegaz, T., Small, L., Leflore, D., Williams, J., Adams, J., Dozi‐ er, G., Bryant, K., Ricanek, K., & Woodard, D. L. (2012). Genetic & Evolutionary Bio‐ metric Security: Feature Extraction from a Machine Learning Perspective. *IEEE*

[27] Shelton, J., Venable, M., Neal, S., Adams, J., Alford, A., & Dozier, G. (2012). *Pixel Con‐ sistency, K-Tournament Selection, and Darwinian-Based Feature Extraction*, *Midwest Arti‐*

[28] Adams, J., Shelton, J., Small, L., Venable, M., Neal, S., Kim, J., & Dozier, G. (2012). Darwinian-based Feature Extraction using K-means and Kohonen Clustering. Paper presented at Midwest Artificial Intelligence and Cognitive Science (MAICS) Confer‐

[29] Shelton, J., Dozier, G., Bryant, K., Smalls, L., Adams, J., Popplewell, K., Abegaz, T., Woodard, D., & Ricanek, K. (2011). Genetic and Evolutionary Feature Extraction via X-TOOLSS. in The 8th Annual International Conference on Genetic and Evolutionary

[30] Alford, A., Bryant, K., Abegaz, T., Dozier, G., Kelly, J., Shelton, J., Small, L., Williams, J., & Woodard, D. L. (2011). Genetic & Evolutionary Methods for Biometric Feature Reduction. Special Issue on: "Computational Intelligence in Biometrics: Theory, Methods and Applications", Guest Editor: Qinghan Xiao, International Journal of

[31] Alford, A., Adams, J., Shelton, J., Bryant, K., Kelly, J. C., & Dozier, G. (2012). Analyz‐ ing the Cross-Generalization Ability of a Hybrid Genetic & Evolutionary Application for Multibiometric Feature Weighting and Selection. *Genetic and Evolutionary Compu‐*

[32] Ojala, T., Pietikainen, M., & Harwood, D. (1996). A comparative study of texture measures with classification based on feature distributions. *Pattern Recognition*, 29,

[33] Ojala, T., Pietikainen, M., & Mäenpää, T. (2002). MultiresolutionGray-scale and Rota‐ tion Invariant Texture Classification with Local Binary Patterns. *Proceedings of IEEE*

[34] Marcel, S., Rodriguez, Y., & Heusch, G. (2006). On The Recent Use of Local Binary Patterns for Face Authentication. *International Journal of Image and Video Processing-*

*Trans. Pattern Analysis and Machine Intelligence*, 24(7), 971-987.

*ficial Intelligence and Cognitive Science (MAICS) Conference*.

*Berlin /Heidelberg*, 317-330.

*SoutheastCon 2012*.

Methods (GEM).

Biometrics.

51-59.

*tation Conference (GECCO-2012)*.

*Special Issue on Facial Image Processing*, 1-9.

ence.

System. *IEEE SoutheastCon 2012*.


[24] Dahal, K. P., Tan, K. C., & Cowling, P. I. (2007). Evolutionary Scheduling. *Springer Berlin /Heidelberg*, 317-330.

[10] Shelton, J., Dozier, G., Adams, J., & Alford, A. (2012). Permutation-Based Biometric Authentication Protocols for Mitigating Replay Attacks. Paper presented at Proceed‐

ings of the 2012 IEEE World Congress on Computational Intelligence (WCCI).

*Symposium on Computational Intelligence for Security and Defense Applications.*

[11] Shelton, J., Adams, J., Alford, A., Venable, M., Neal, S., Dozier, G., & Bryant, K. (2012). Mitigating Replay Attacks Using Darwinian-Based Feature Extraction. *IEEE*

[12] Goldberg, D. E. (1989). Genetic Algorithms in Search, Optimization & Machine Learning. Addison-Wesley Publishing Company, Inc., Reading, Massachusetts.

[13] Kennedy, J., Eberhart, R. C., & Shi, Y. (2001). *Swarm Intelligence*, Morgan Kaufmann.

[14] Nolfi, S., & Floreano, D. (2000). Evolutionary Robotics:The Biology, Intelligence and

[15] Vadakkepat, P., Chen, Tan. K., & Ming-Liang, W. (2000). Evolutionary artificial po‐ tential fields and their application in real time robot path planning. *Proceedings of the*

[16] Hoffmann, F., & Pfister, G. (1997, November). Evolutionary design of a fuzzy knowl‐ edge base for a mobile robot. *International Journal of Approximate Reasoning*, 17(4),

[17] Harvey, I., Husbands, P., Cliff, D.., Thompson, A., & Jakobi, N. (1997, June). Evolu‐ tionary robotics: the Sussex approach. *Robotics and Autonomous Systems*, 20(2-4),

[18] Bentley, P. J. (1999). Evolutionary design by computers. San Francisco, Morgan Kauf‐

[19] Gargiulo, A. A., Schnell, A. R., Tinker, M. L., Dozier, G., & Garrett, A. (2000). Use of Evolutionary Computation for Three-Dimensional Structural Design of a Lunar Hab‐ itat. http://nxt.ncat.edu/pdf/Use%20of%20EC\_for\_Prelim\_Struct\_Design\_Lunar\_

[20] Dozier, G., Carnahan, B., Seals, C., Kuntz, L. A., & Ser-Geon, Fu. (2005). An interac‐ tive distributed evolutionary algorithm (IDEA) for design. 2005 IEEE International

[21] Miranda, E. R. (2004). At the Crossroads of Evolutionary Computation and Music: Self-Programming Synthesizers, Swarm Orchestras and the Origins of Melody. *Evo‐*

[22] Murata, T., Ishibuchi, H., & Tanaka, H. (1996, September). Multi-objective genetic al‐ gorithm and its applications to flowshop scheduling. *Computers & Industrial Engineer‐*

[23] Hart, E., Ross, P., & Corne, D. (2005). Evolutionary Scheduling: A Review. *Genetic*

Conference on Systems, Man and Cybernetics, 10-12 Oct , 1, 418-422.

Technology of Self-Organizing Machines. MIT Press.

*2000 Congress on Evolutionary Computation*, 256-263, 1.

447-469.

122 New Trends and Developments in Biometrics

205-224.

mann.

Habitat.pdf.

*ing*, 30(4), 957-968.

*lutionary Computation*, 12(2), 137-158.

*Programming and Evolvable Machines*, 6(2), 191-220.


[35] Galbally, J., Fierrez, J., Freire, M., & Ortega-Garcia, J. (2007). Feature Selection Based on Genetic Algoirthms for On-Line Signature Verification. Paper presented at IEEE Workshop on Automatic Identification Advanced Technologies Alghero, Italy. 198-203.

Paper presented at Proceedings of the 1999 IEEE Signal Processing Society Work‐

Genetic & Evolutionary Biometrics http://dx.doi.org/10.5772/51386 125

[49] Schölkopf, B., Smola, A., & Müller, K. R. (1997). Kernel Principal Component Analy‐ sis. *ARTIFICIAL NEURAL NETWORKS- ICANN'97, Lecture Notes in Computer Science*,

[50] Larrañaga, P., & Lozano, J. A. (2002). Estimation of Distribution Algorithms: A new

[51] Turk, M., & Pentland, A. (1991). Eigenfaces for recognition. *Journal of Cognitive Neuro‐*

[52] Phillips, P. J., Flynn, P. J., Scruggs, T., Bowyer, K. W., Chang, J., Hoffman, K., Mar‐ ques, J., Min, J., & Worek, W. (2005). Overview of face recognition grand challenge. Paper presented at Proc. IEEE Conference on Computer Vision and Pattern Recogni‐

[53] Ricanek, K., & Tesafaye, T. (2006, April 10-12.) MORPH: A Longitudinal Image Data‐ base of Normal Adult Age-Progression". Paper presented at Proceedings of the 7th International Conference on Automatic Face and Gesture Recognition. 341-345.

[55] Kanungo, T., Mount, D., Netanyahu, N., Piatko, C., Silverman, R., & Wu, A. (2002, July). An Efficient k-Means Clustering Algorithm: Analysis and Implementation.

[57] Specht, D. F. (1991, Nov). A general regression neural network. *IEEE Transactions on*

[58] Miller, P. E., Rawls, A. W., Pundlik, S. J., & Woodard, D. L. (2010). Personal identifi‐ cation using periocular skin texture. Proceedings of the 2010 ACM Symposium on

[59] Srensen, L., Shaker, S. B., & de Bruijne, M. (2010, Feb). Quantitative analysis of pul‐ monary emphysema using local binary patterns. *IEEE Transactions on Medical Imag‐*

[60] Unay, D., Ekin, A., Cetin, M., Jasinschi, R., & Ercil, A. (2007, August 23-26,). Robust‐ ness of Local Binary Patterns in Brain MR Image Analysis. Paper presented at Pro‐ ceedings of the 29th Annual International Conference of the IEEE Engineering in

[61] Woodard, D., Pundlik, S., Lyle, J., & Miller, P. (2010). Periocular region appearance cues for biometric identification. CVPR Workshop on Biometrics. San Francisco, CA ,

[54] Mitchell, T. M. (1997). *Machine Learning*, McGraw-Hill Companies, Inc.

*IEEE Transactions on Pattern Analysis and Machine Intelligence*, 24(7).

[56] Kohonen, T. (1990). The self-organising map. *Proceedings IEEE*, 78(9), 1464-1480.

Applied Computing (SAC'10). ACM, New York, NY, USA, 1496-1500.

Medicine and Biology Society (EMBS), Lyon, France,. 2098-2101.

shop. 41-48.

1327, 583-588.

*science*, 13(1), 71-86.

*Neural Networks*, 2(6), 568-576.

*ing*, 29(2), 559-569.

162-169.

tion.

tool for evolutionary computation. Springer.


Paper presented at Proceedings of the 1999 IEEE Signal Processing Society Work‐ shop. 41-48.

[49] Schölkopf, B., Smola, A., & Müller, K. R. (1997). Kernel Principal Component Analy‐ sis. *ARTIFICIAL NEURAL NETWORKS- ICANN'97, Lecture Notes in Computer Science*, 1327, 583-588.

[35] Galbally, J., Fierrez, J., Freire, M., & Ortega-Garcia, J. (2007). Feature Selection Based on Genetic Algoirthms for On-Line Signature Verification. Paper presented at IEEE Workshop on Automatic Identification Advanced Technologies Alghero, Italy.

[36] Ramadan, R. M., & Abdel-Kader, R. F. (2009, June). Face Recognition Using Particle Swarm Optimization-Based Selected Features. *International Journal of Signal Process‐*

[37] Abegaz, T., Dozier, G., Bryant, K., Adams, J., Mclean, V., Shelton, J., Alford, A., Rica‐ nek, K., & Woodard, D. L. (2011). Applying GECs for Feature Selection and Weight‐ ing using X-TOOLSS". *The 8th Annual International Conference on Genetic and*

[38] Alford., A., Popplewell, K., Dozier, G., Bryant, K., Kelly, J., Adams, J., Abegaz, T., Shelton, J., Woodard, D. L., & Ricanek, K. (2011). Hybrid GEMs for Multi-Biometric Recognition via X-TOOLSS. *The 8th Annual International Conference on Genetic and Evo‐*

[39] Adams, J., Dozier, G., Bryant, K., Shelton, J., Alford, A., Abegaz, T., & Leflore, D. (2012, 15-18 March). Neurogenetic Reconstruction of Biometric Templates: A New

[40] Garcia- Ortega, J., Fierrez-Aguilar, J., Simon, D., Gonzalez, J., Faundez-Zanuy, M., Espinosa, V., Satue, A., Hernaez, I., Igarza, J. J., Vivaracho, C., Escudero, C., & Moro, Q. I. (2003, December). MCYT baseline corpus: a bimodal biometric database. *IEEE*

[41] AT&T Laboratories Cambridge. (2000). ORL Face Database. http://

[42] Ahmed, N., Natarajan, T., & Rao, K. R. (1974, Jan). Discrete Cosine Transform. *IEEE*

[43] Shensa, M. J. (1992, Oc). The Discrete Wavelet Transform: Wedding the À Trous and Mallat Algorithms. *IEEE Transactions on Signal Processing*, 40(10), 2464-2482.

[44] Kumar, D., Kumar, S., & Rai, C. S. (2009). Feature selection for face recognition: a memetic algorithmic approach. *Journal of Zhejanga University Science A*, 10(8),

[45] Neri, F., Cotta, C., & Moscato, P. (2012). Handbook of Memetic Algorithms. *Studies in*

[46] PAMI. (2001). The Extended Yale Face Database B. http://vision.ucsd.edu/~leekc/

[48] Mika, S., Ratsch, G., Weston, J., Scholkopf, B., & Mullers, K. R. (1999, Aug). Fisher discriminant analysis with kernels," Neural Networks for Signal Processing IX, 1999.

[47] Jollife, I. T. (1986). *Principal Component Analysis.*, New York, Springer-Verlag.

Security Threat? Paper presented at IEEE Southeast Con 2012, Orlando, FL.

*ing, Image Processing and Pattern Recognition*, 2(2).

*Proc. Vis. Image Signal Process.*, 150(6), 395-401.

*Transactions on Computers*, C-23(1), 90-93.

*Computational Intelligence*, Springer, 379.

ExtYaleDatabase/ExtYaleB.html.

1140-1152.

www.cl.cam.ac.uk/research/dtg/attarchive/facedatabase.html.

*Evolutionary Methods(GEM)*.

*lutionary Methods (GEM).*

198-203.

124 New Trends and Developments in Biometrics


[62] Miller, B. L., & Goldberg, D. E. (1996, June). Genetic algorithms, selection schemes, and the varying effects of noise. *Evol. Comput*, 4(2), 113-131.

**Section 2**

**Performance Evaluation**
