**3.9 Legal aspects**

Development of 'query by example,'subword, keyword and phrase-based retrieval schemes using conventional and structural similarity-based methods

Documents will be preprocessed for enhancement and noise removal. Exemplarbased word and subword matching will be used to identify and verify the writers. Experts will be provided with a set of results and will be asked for feedback to be

Development of both character and word-based recognition schemes. Manual effort for labeling data during training will be reduced by learning the relationships between available handwritten and printed text pairs. Scene text will also be

The objective is to develop a CBR-based novelty detection method based on a combination of statistical and similarity-based methods able to handle the different

The method will perform novelty detection and handling of the novel cases for immediately reasoning and new model built up and should consider the incremental nature of the data. It will update the models incrementally based on MML and MDL

• State of art on statistical novelty detection methods applicable to multimedia

• Model development for novelty detection classification for multimedia data

Taking into account the outcome of Task 8.1, new methods for the different

The incremental learning mechanism for the updating of the models and feature description should be developed based on statistical learning methods such as MML

Development of the data base structure for the different multimedia data types.

The incremental data collection mechanism will be developed, taking into

• Task management mechanism between the model-based and the similarity-

multimedia data types will be developed and implemented.

• Incremental model learning mechanism based on MDL or MML

• Incremental data collection mechanism and data base structure

Different statistical model methods will be studied and evaluated for the different multimedia data types. The methods that are applicable to the data types

capable of part and whole similarity matching.

• Writer identification

*Digital Forensic Science*

used in an active learning scheme.

• Automatic text recognition

recognized.

methods.

**3.8 Novelty detection**

multimedia data types.

data types.

will be selected.

types

or MDL.

**148**

account the data base structure.

based unit

To provide a clear picture of the current legal framework regulating the process of gathering, processing, analyzing and integrating multimedia data for security and judicial purposes at the European (EU Directives and Regulations) and national level.

To specialize the investigation on legal issues related to the use of sensible data for forensic purposes by the analysis of case studies and EU (ECJ) and national judgments.

To provide a framework of standards, quality indicators and approaches for the preservation and validity assessment of digital evidence for forensic purposes.

To support partners in sorting out potential ethical questions, by supporting the consulting process of the advisory body in solving questions specific to ethics, issues concerning sensible data processing and, generally, rules limiting the use of personal data extracted by massive data processing techniques:


(e.g. authenticity) on which legal validity of evidences can be assessed; Task 2 will be implemented in two steps and with a strong connection with the WP 2, by reference to the classification of typologies of data, formats and specific features.

satisfied. The resulting build architecture will be deployable on a cloud computing infrastructure so that scalability requirements can be tested and shown to meet the

This task will develop several case studies that illustrate the use of the system on

• Novel multimedia data acquisition and automatic analysis methods for forensic multimedia data such as images and videos, text, handwriting, speech and

interpretation of forensic multimedia data comprised of image enhancement

• A forensic case basis in which cases and generalized cases are stored for fast

formerly unseen data and situations and detecting new task for the standard

• A learning unit and a novelty detection unit for dealing with novel and

The toolkit will be used as benchmark for testing instruments and rules

interdependency of phenomena (e.g. environment, health, immigration, crimes prevention, anti-terrorism fight, etc.) require law makers and courts to face the evolution of social phenomena and their legal effects within a pan-European harmonized area of justice. European judges tend to adjust the national legal framework by referring to common and shared principles, thus invoking constitutional rules as higher principles to which anchor case solutions. The problem of adopting standards of conduct in judicial procedures seems no longer to be a questionable issue but rather looks like an undeniable fact at the center of a vivid discussion within the international community of jurists, scholars and courts. In a field of

The scenario: the achievement of an effective set of technical solutions and of (law-compliant) procedural standards is a valuable goal for the 'global society',

• New standards and a new methodology for the analysis of forensic

where cultural interaction, de-territorialization of social behaviors and

• A novel toolkit for the automatic and semi-automatic analysis and

real but anonymized data from the law enforcement agencies and end-user companies. These case studies will illustrate the use of video and image enhancement filtering and assessment (PMMW images), case-based reasoning (data provided by all data providers), multimedia feature extraction (data provided by all data providers), video analysis, handwriting recognition, text analysis and novelty detection for forensic analysis of multimedia data (data provided by all data providers). The case studies will also show how the system complies with legal requirements. The case studies will be incrementally developed through the

As the expected result, we will have the following:

*Novel Methods for Forensic Multimedia Data Analysis: Part II*

*DOI: http://dx.doi.org/10.5772/intechopen.92548*

audio signals and social media data

involving technical skills, operators and courts.

requirements.

• Case applications

different project iterations.

**4. Expected results**

algorithm

retrieval and reasoning

multimedia data.

**151**


### **3.10 Evaluation**

To integrate and test the different components in an application that can easily be deployed in a computing infrastructure. This task aims to perform integration testing of the components and building an easily deployable application.

To develop case studies that will illustrate the features of the system. The case studies will illustrate the need for video and image enhancement, filtering and assessment, case-based reasoning, multimedia feature extraction, video analysis, handwriting recognition and novelty detection and illustrate how the system complies with legal requirements.

To evaluate the developed methods in the integrated software solution and to demonstrate the applicability of the solutions.

Risk mapping on territory, or based on the results of the scenarios, to determine certain potentially illegal behavior occurrences involving the activation of protection and security, both in terms of allocation of armed guards and installation of dedicated electronic equipment;

Support the decision on whether to carry out an intervention and which mode to adopt, on the basis of the evolution of a given behavioral scenario.


So the solution needs to be evaluated against real life situations.

• Integration and testing

This task will integrate the components produced by the different tasks. This will require integrating the video and image components, the case-based reasoning components, the multimedia feature extraction component, the social media text analytics component, the video analysis and face detection component, the speech and audio processing component, the handwriting recognition component and the novelty detection component. The system build will then undergo integration and testing to ensure that architectural functional and nonfunctional requirements are

satisfied. The resulting build architecture will be deployable on a cloud computing infrastructure so that scalability requirements can be tested and shown to meet the requirements.

• Case applications

(e.g. authenticity) on which legal validity of evidences can be assessed; Task 2 will be implemented in two steps and with a strong connection with the WP 2, by reference to the classification of typologies of data, formats and specific

acquiring, processing, storing and preserving digital evidences, based on the selection of guidelines and best practices of relevant judicial cases and of significant literature. (i.e., ENFSI community (http://www.enfsi.org/))

• A**nalysis of the communication process** between the judge, who is the final evaluator ('free conviction of the court') and the expert, who plays the role of a

To integrate and test the different components in an application that can easily be deployed in a computing infrastructure. This task aims to perform integration

To develop case studies that will illustrate the features of the system. The case studies will illustrate the need for video and image enhancement, filtering and assessment, case-based reasoning, multimedia feature extraction, video analysis, handwriting recognition and novelty detection and illustrate how the system com-

To evaluate the developed methods in the integrated software solution and to

Risk mapping on territory, or based on the results of the scenarios, to determine certain potentially illegal behavior occurrences involving the activation of protection and security, both in terms of allocation of armed guards and installation of

Support the decision on whether to carry out an intervention and which mode to

• **Identification** of a large class of—potentially or effectively—illegal behaviors.

This task will integrate the components produced by the different tasks. This will require integrating the video and image components, the case-based reasoning components, the multimedia feature extraction component, the social media text analytics component, the video analysis and face detection component, the speech and audio processing component, the handwriting recognition component and the novelty detection component. The system build will then undergo integration and testing to ensure that architectural functional and nonfunctional requirements are

• **Registration and selective search** of details from an offense scenario

So the solution needs to be evaluated against real life situations.

testing of the components and building an easily deployable application.

adopt, on the basis of the evolution of a given behavioral scenario.

• **Analysis of technical and legal requirements** (e.g. ISO Guide 25) for

'mediator', expected to provide the basic elements (time, location, authenticity, etc.) on which the evidence is grounded. The cognitive background, the communicative interaction, the semantic/terminological mapping and the reasoning processes will be modeled in an integrated picture; on this topics Task 9.2 will publish a report that aims at providing the scientific community with an update and original approach to the methodology for

quality assessment of digital evidence.

demonstrate the applicability of the solutions.

features.

*Digital Forensic Science*

**3.10 Evaluation**

plies with legal requirements.

dedicated electronic equipment;

• Integration and testing

**150**

This task will develop several case studies that illustrate the use of the system on real but anonymized data from the law enforcement agencies and end-user companies. These case studies will illustrate the use of video and image enhancement filtering and assessment (PMMW images), case-based reasoning (data provided by all data providers), multimedia feature extraction (data provided by all data providers), video analysis, handwriting recognition, text analysis and novelty detection for forensic analysis of multimedia data (data provided by all data providers). The case studies will also show how the system complies with legal requirements. The case studies will be incrementally developed through the different project iterations.
