**3. Proposed method**

The characteristic value of the target pixel extracted by LGS and MOW-SLGS algorithm are in the neighborhood of 3 × 4 and 5 × 5. It can be seen that these two algorithms do not consider all the surrounding pixels in the neighborhood. This leads to the lack of information about the surrounding pixels, which will have an impact on the recognition rate.

In **Figure 5**, we calculate the characteristic value according to our proposed LCGS algorithm. The target pixel value is 22. The value of target pixel is 22. We compare the value of pixels in according to the order of X<sup>0</sup> → X<sup>8</sup> → X<sup>1</sup> → X<sup>0</sup> → X<sup>2</sup> → X3 → X<sup>0</sup> → X<sup>4</sup> → X5 → X<sup>0</sup> → X<sup>6</sup> → X<sup>7</sup> → X<sup>0</sup>

Face Recognition with Facial Occlusion Based on Local Cycle Graph Structure Operator

http://dx.doi.org/10.5772/intechopen.78597

and get the binary values as 001010101101. It is converted to the decimal number, that is, F (X<sup>0</sup>

=0 × 2048 + 0 × 1024 + 1 × 512 + 0 × 256 + 1 × 128 + 0 × 64 + 1 × 32 + 0 × 16 + 1 × 8 + 1 × 4 + 0 × 2 +

In the LCGS algorithm, we define the neighborhood of 3 × 3 along the traditional way. It is effective to calculate the characteristic values of the target pixel by considering all the surrounding pixels in the neighborhood. Thus, it makes the final recognition rate improved.

In order to investigate the reliability of LCGS algorithm, we use the AR physical occlusion database and the ORL simulation occlusion database to conduct the experiments. The AR face database consists of 126 people, a total of 3276 images, 26 pictures per person. The 26 pictures are collected in two different time periods, including changes in light, occlusion, and so on. The shelter is divided into scarves and sunglasses. The ORL Face Database was created by the University of Cambridge AT &T Lab, which contains 40 people, 10 images per person, and a total of 400 face images. We use a baboon image to randomly block the original image for the occlusion simula-

tion. Some examples of AR database and ORL database are shown in **Figures 6** and **7**.

1 × 1 = 685, which gives the final feature values for the target pixel.

**4. Experiments and analysis**

**Figure 5.** Template of LCGS operator.

**Figure 4.** LCGS operator.

,

61

)

In order to solve this problem, we propose a LCGS algorithm by using the full information of the surrounding pixels. In such way, LCGS can get more representative features for the target pixels, thus, improve the recognition rate.

The way to extract the characteristics of the face image with LCGS algorithm is illustrated in **Figure 4**.

In **Figure 4**, the center pixel X<sup>0</sup> is taken as the target pixel in the neighborhood of 3 × 3. The remaining pixels are represented by X<sup>1</sup> to X8, respectively. We put all the pixels in accordance with X<sup>0</sup> → X<sup>8</sup> → X<sup>1</sup> → X<sup>0</sup> , X<sup>0</sup> → X<sup>2</sup> → X3 → X<sup>0</sup> , X<sup>0</sup> → X<sup>4</sup> → X5 → X<sup>0</sup> , X<sup>0</sup> → X<sup>6</sup> → X<sup>7</sup> → X<sup>0</sup> to compare the value of the pixel. If the value of the pixel is larger in the direction of the arrow, we set it to 1; otherwise, we set it to 0. Finally, we obtain 12 binary numbers in sequence from X<sup>0</sup> → X<sup>8</sup> to X<sup>7</sup> → X<sup>0</sup> . The decimal format of these 12 binary values is the characteristic value. A specific example is shown in **Figure 5**.

Face Recognition with Facial Occlusion Based on Local Cycle Graph Structure Operator http://dx.doi.org/10.5772/intechopen.78597 61

**Figure 4.** LCGS operator.

**Figure 5.** Template of LCGS operator.

**Figure 3.** Graph structure of MOW-SLGS in the direction of 45°.

**Figure 2.** The design of weights for MOW-SLGS.

60 Machine Learning and Biometrics

pixels, thus, improve the recognition rate.

In **Figure 4**, the center pixel X<sup>0</sup>

example is shown in **Figure 5**.

with X<sup>0</sup> → X<sup>8</sup> → X<sup>1</sup> → X<sup>0</sup>

remaining pixels are represented by X<sup>1</sup>

The characteristic value of the target pixel extracted by LGS and MOW-SLGS algorithm are in the neighborhood of 3 × 4 and 5 × 5. It can be seen that these two algorithms do not consider all the surrounding pixels in the neighborhood. This leads to the lack of information about the

In order to solve this problem, we propose a LCGS algorithm by using the full information of the surrounding pixels. In such way, LCGS can get more representative features for the target

The way to extract the characteristics of the face image with LCGS algorithm is illustrated in

the value of the pixel. If the value of the pixel is larger in the direction of the arrow, we set it to 1; otherwise, we set it to 0. Finally, we obtain 12 binary numbers in sequence from X<sup>0</sup> → X<sup>8</sup>

, X<sup>0</sup> → X<sup>4</sup> → X5 → X<sup>0</sup>

. The decimal format of these 12 binary values is the characteristic value. A specific

is taken as the target pixel in the neighborhood of 3 × 3. The

to X8, respectively. We put all the pixels in accordance

, X<sup>0</sup> → X<sup>6</sup> → X<sup>7</sup> → X<sup>0</sup>

to compare

surrounding pixels, which will have an impact on the recognition rate.

, X<sup>0</sup> → X<sup>2</sup> → X3 → X<sup>0</sup>

**3. Proposed method**

**Figure 4**.

to X<sup>7</sup> → X<sup>0</sup>

In **Figure 5**, we calculate the characteristic value according to our proposed LCGS algorithm. The target pixel value is 22. The value of target pixel is 22. We compare the value of pixels in according to the order of X<sup>0</sup> → X<sup>8</sup> → X<sup>1</sup> → X<sup>0</sup> → X<sup>2</sup> → X3 → X<sup>0</sup> → X<sup>4</sup> → X5 → X<sup>0</sup> → X<sup>6</sup> → X<sup>7</sup> → X<sup>0</sup> , and get the binary values as 001010101101. It is converted to the decimal number, that is, F (X<sup>0</sup> ) =0 × 2048 + 0 × 1024 + 1 × 512 + 0 × 256 + 1 × 128 + 0 × 64 + 1 × 32 + 0 × 16 + 1 × 8 + 1 × 4 + 0 × 2 + 1 × 1 = 685, which gives the final feature values for the target pixel.

In the LCGS algorithm, we define the neighborhood of 3 × 3 along the traditional way. It is effective to calculate the characteristic values of the target pixel by considering all the surrounding pixels in the neighborhood. Thus, it makes the final recognition rate improved.
