Contents. 3 Improving Face Recognition Using Directional Faces Introduction xiii

Transcription

1 Contents 1 Introduction and Preliminaries on Biometrics and Forensics Systems Introduction Definition of Biometrics BiometricCharacteristics Biometric Modalities Recognition/Verification/Watch-List Verification:AmIWhoIClaimtoBe? Recognition: Who Am I? The Watch-List: Are You Looking for Me? Steps of a Typical Biometric Recognition Application BiometricDataLocalisation NormalisationandPre-processing Feature Extraction Matching Databases Summary... 9 References Data Representation and Analysis Introduction Data Acquisition Sensor Module DataStorage Feature Extraction Matcher SystemTesting Performance Evaluation Conclusion References Improving Face Recognition Using Directional Faces Introduction xiii

2 xiv Contents 3.2 Face Recognition Basics Recognition/Verification Steps of a Typical Face Recognition Application PreviousWork Principal Component Analysis (PCA) Independent Component Analysis (ICA) Linear Discriminant Analysis (LDA) Subspace Discriminant Analysis (SDA) Face Recognition Using Filter Banks Gabor Filter Bank Directional Filter Bank: A Review Proposed Method and Results Analysis Proposed Method PCA ICA LDA SDA FERET Database Results Conclusion References Recent Advances in Iris Recognition: A Multiscale Approach Introduction RelatedWork:AReview IrisLocalisation Background IrisSegmentation Existing Methods for Iris Localisation Proposed Method for Iris Localisation Motivation The Multiscale Method ResultsandAnalysis Texture Analysis and Feature Extraction Wavelet Maxima Components Special Gabor Filter Bank Proposed Method Matching Experimental Results and Analysis Database Combined Multiresolution Feature Extraction Techniques TemplateComputation Comparison with Existing Methods DiscussionandFutureWork Conclusion References

3 Contents xv 5 Spread Transform Watermarking Using Complex Wavelets Introduction WaveletTransforms DualTreeComplexWaveletTransform Non-redundant Complex Wavelet Transform Visual Models Chou s Model Loo s Model Hybrid Model WatermarkingasCommunicationwithSideInformation Quantisation Index Modulation SpreadTransformWatermarking Proposed Algorithm Encoding of Watermark Decoding of Watermark Information Theoretic Analysis Decoding of Watermark Parallel Gaussian Channels WatermarkingGame Non-iidData Fixed Embedding Strategies Conclusion References Protection of Fingerprint Data Using Watermarking Introduction Generic Watermarking System State-of-the-Art OptimumWatermarkDetection Statistical Data Modelling and Application to Watermark Detection Laplacian and Generalised Gaussian Models Alpha Stable Model Experimental Results Experimental Modelling of DWT Coefficients Experimental Watermarking Results Conclusions References Shoemark Recognition for Forensic Science: An Emerging Technology Background to the Problem of Shoemark Forensic Evidence Applications of a Shoemark in Forensic Science The Need for Automating Shoemark Classification Inconsistent Classification

4 xvi Contents Importable Classification Schema Shoemark Processing Time Restrictions Collection of Shoemarks at Crime Scenes Shoemark Collection Procedures Transfer/Contact Shoemarks Photography of Shoemarks Making Casts of Shoemarks Gelatine Lifting of Shoemarks Electrostatic Lifting of Shoemarks Recovery of Shoemarks from Snow Recovery of Shoemarks using Perfect Shoemark Scan Making a Cast of a Shoemark Directly from a Suspect s Shoe Processing of Shoemarks EnteringDataintoaComputerisedSystem Typical Methods for Shoemark Recognition Feature-Based Classification Classification Based on Accidental Characteristics Review of Shoemark Classfication Systems SHOE-FIT SHOE Alexandre s System REBEZO TREADMARK TM SICAR SmART De Chazal s System Zhang s System References Techniques for Automatic Shoeprint Classification Current Approaches Using Phase-Only Correlation The POC Function Translation and Brightness Properties of the POC Function The Proposed Phase-Based Method Experimental Results DeploymentofACFs Shoeprint Classification Using ACFs MatchingMetrics Optimum Trade-Off Synthetic Discriminant Function Filter Unconstrained OTSDF Filter TestsandResults...178

5 Contents xvii 8.4 Conclusion References Automatic Shoeprint Image Retrieval Using Local Features Motivations Local Image Features New Local Feature Detector: Modified Harris Laplace Detector Local Feature Descriptors SimilarityMeasure Experimental Results Shoeprint Image Databases Summary References Index...203

6