Cover
Half Title
Series Page
Title Page
Copyright Page
Table of Contents
Preface
Editors
Contributors
Chapter 1 Deep Learning-Based Hyperspectral Multimodal Biometric Authentication System Using Palmprint and Dorsal Hand Vein
1.1 Introduction
1.2 Device Design
1.3 System Implementation
1.3.1 ROI Extraction
1.3.1.1 Hyperspectral Palmprint ROI Extraction
1.3.1.2 Hyperspectral Dorsal Hand Vein ROI Extraction
1.3.2 Feature Extraction
1.3.3 Feature Fusion and Matching
1.4 Experiments
1.4.1 Multimodal Hyperspectral Palmprint and Dorsal Hand Vein Dataset
1.4.2 Optimal Pattern and Band Selection
1.4.3 Multimodal Identification
1.4.4 Multimodal Verification
1.4.5 Computational Complexity Analysis
1.5 Conclusions
Acknowledgements
References
Chapter 2 Cancelable Biometrics for Template Protection: Future Directives with Deep Learning
2.1 Introduction
2.2 Template Protection
2.2.1 Consequences of Template Compromise
2.2.2 Template Protection Techniques
2.2.3 Comparative Analysis between Template Protection Techniques
2.2.4 Fundamental Requirements of Template Protection Techniques
2.2.5 Potential Attacks on Template Protection Techniques
2.3 Role of Deep Learning Approaches in Biometrics
2.3.1 Deep Learning in Face Recognition
2.3.2 Deep Learning in Iris Recognition
2.3.3 Deep Learning in Fingerprint Recognition
2.3.4 Deep Learning in Other Biometric Traits
2.4 Related Work: Template Protection
2.4.1 Biometric Encryption
2.4.2 Biometric Cryptosystems
2.4.3 Cancelable Biometrics
2.4.3.1 Deep Learning-Based Cancelable Techniques
2.4.3.2 Deep Learning versus Non-deep Learning Cancelable Techniques
2.5 Performance Measures and Datasets in Cancelable Biometrics
2.5.1 Performance Measures for Non-invertibility Analysis
2.5.2 Performance Measures for Unlinkability Analysis
2.5.3 Performance Measures for System Usability Analysis
2.5.4 Performance Measures for Revocability Analysis
2.5.5 Databases Used in Cancelable Biometrics
2.6 Comparative Performance Analysis: Cancelable Biometrics
2.7 Conclusions and Future Prospective of Deep Learning in Biometrics
References
Chapter 3 On Training Generative Adversarial Network for Enhancement of Latent Fingerprints
3.1 Introduction
3.2 Related Work
3.3 Proposed Algorithm
3.3.1 Problem Formulation and Objective Function
3.3.2 Training Data Preparation
3.3.3 Network Architecture and Training Details
3.4 Performance Evaluation
3.4.1 Databases and Tools Used
3.4.2 Evaluation Criteria
3.5 Results and Analysis
3.6 Challenges Observed
3.7 Conclusions
Acknowledgements
References
Chapter 4 DeepFake Face Video Detection Using Hybrid Deep Residual Networks and LSTM Architecture
4.1 Introduction
4.2 Related Work
4.2.1 Categories of Face Manipulations
4.2.1.1 Face Synthesis
4.2.1.2 Face Swap
4.2.1.3 Facial Attributes
4.2.1.4 Face Expression
4.2.2 DeepFakes Detection
4.3 Proposed DeepFake Videos Detection Framework
4.3.1 Convolutional Neural Networks (CNNs)
4.3.2 Long Short-Term Memory (LSTM)
4.3.3 Residual Neural Network (ResNet)
4.4 Experiments
4.4.1 Datasets
4.4.1.1 DeepFakeTIMIT Dataset
4.4.1.2 Celeb-DF Dataset
4.4.2 Figures of Merit
4.4.3 Experimental Protocol
4.4.4 Experimental Results
4.5 Challenges and Future Research Directions
4.6 Conclusions
Notes
References
Chapter 5 Multi-spectral Short-Wave Infrared Sensors and Convolutional Neural Networks for Biometric Presentation Attack Detection
5.1 Introduction
5.2 Definitions
5.3 Related Works
5.4 Proposed PAD Method
5.4.1 Hardware: Multi-Spectral SWIR Sensor
5.4.2 Software: Multi-Spectral Convolutional Neural Networks
5.4.2.1 Multi-Spectral Samples Pre-Processing
5.4.2.2 CNN Models
5.4.2.3 Score Level Fusion
5.5 Experimental Setup
5.5.1 Database
5.5.2 Evaluation Metrics
5.5.3 Experimental Protocol
5.6 Experimental Evaluation
5.6.1 Baseline: Handcrafted RGB Conversion
5.6.2 Input Pre-Processing Optimisation
5.6.3 Final Fused System
5.7 Conclusions and Future Research
Acknowledgements
References
Chapter 6 AI-Based Approach for Person Identification Using ECG Biometric
6.1 Introduction
6.2 ECG and Related Work
6.2.1 Advantages of ECG Biometric
6.2.2 Literature Review
6.3 Methodology Adopted
6.3.1 Feature Extraction
6.4 Classifier
6.4.1 Artificial Neural Network (ANN)
6.4.2 Support Vector Machine (SVM)
6.5 Experiments and Results
6.6 Conclusions
References
Chapter 7 Cancelable Biometric Systems from Research to Reality: TheΒ Road Less Travelled
7.1 Introduction
7.2 Cancelable Biometric Systems: Introduction and Review
7.2.1 Conventional Template Transformation Techniques
7.2.2 Role of Deep Learning in Biometrics and Need for Privacy
7.2.3 Neutral Network-Based Template Transformation Techniques
7.3 Experimental Reporting
7.4 Real-Life Challenges for Applications of Cancelable Biometric Systems
7.5 Conclusions and Foresights
References
Chapter 8 Gender Classification under Eyeglass Occluded Ocular Region: An Extensive Study Using Multi-spectral Imaging
8.1 Introduction
8.1.1 Our Contributions
8.2 Related Works
8.2.1 Visible Spectrum
8.2.2 Near-Infra-Red Spectrum
8.2.3 Visible and Near-Infra-Red Spectrum
8.2.4 Multi-Spectral Imaging
8.3 Database
8.3.1 Data Preprocessing
8.4 Proposed Method
8.4.1 Spectral Bands Selection
8.4.2 Feature Extraction
8.4.3 Classification
8.5 Experiments and Results
8.5.1 Experimental Evaluation Protocol
8.5.2 Evaluation 1: Without-Glass v/s Without-Glass
8.5.2.1 Individual Band Comparison
8.5.2.2 Fused Band Comparison
8.5.3 Evaluation 2: Without-Glass v/s With-Glass
8.5.3.1 Individual Band Comparison
8.5.3.2 Fused Band Comparison
8.6 Conclusions
Acknowledgement
References
Chapter 9 Investigation of the Fingernail Plate for Biometric Authentication using Deep Neural Networks
9.1 Introduction
9.1.1 Motivation and Scope of Present Work
9.2 Related Work
9.3 Sample Acquisition and ROI Extraction
9.3.1 Sample Acquisition
9.3.2 ROI Extraction
9.4 Feature Extraction
9.4.1 Transfer Learning using AlexNet
9.4.2 Transfer Learning using ResNet-18
9.4.3 Transfer Learning using DenseNet-201
9.5 Multimodal System Design
9.5.1 Score-Level Fusion
9.5.2 Rank-Level Fusion
9.5.2.1 Logistic Regression Method
9.5.2.2 Mixed Group Rank
9.5.2.3 Inverse Rank Position
9.5.2.4 Nonlinear Weighted Methods
9.6 Experiments, Results, and Analyses
9.6.1 Performance of Fingernail Plates in Verification Systems
9.6.1.1 Performance of Fingernail Plates in Unimodal Verification Systems
9.6.1.2 Performance of Fingernail Plates in Multimodal Verification Systems
9.6.2 Performance of Fingernail Plates in Identification Systems
9.6.2.1 Performance of Fingernail Plates in Unimodal Identification Systems
9.6.2.2 Performance of Fingernail Plates in Multimodal Identification Systems
9.7 Challenges and Scope of Fingernail Plates in Biometrics
9.8 Conclusions and Future Scope
References
Chapter 10 Fraud Attack Detection in Remote Verification Systems for Non-enrolled Users
10.1 Introduction
10.2 Related Work
10.2.1 Remote Authentication Framework Using Biometrics
10.2.2 Image Manipulation and Deep Learning Techniques
10.3 Fake ID Card Detection for Non-enrolled Users
10.3.1 Databases
10.3.2 Hand-Crafted Feature Extraction (BSIF, uLBP, and HED)
10.3.3 Automatic Feature Extraction (CNN)
10.4 Experiments and Results
10.4.1 Feature Extraction Classification
10.4.2 Classification Using CNN Algorithms
10.4.2.1 Small-VGG Trained from Scratch
10.4.2.2 Pre-trained VGG16 Model and Bottleneck
10.4.2.3 Pre-trained VGG16 Model and Fine-Tuning
10.5 Conclusions
Acknowledgement
References
Chapter 11 Indexing on Biometric Databases
11.1 Introduction
11.2 Indexing Facial Images
11.2.1 Predictive Hash Code
11.2.2 Results
11.3 Indexing Fingerprint Images
11.3.1 Coaxial Gaussian Track Code
11.3.2 Results
11.4 Indexing Finger-Knuckle Print Database
11.4.1 Boosted Geometric Hashing
11.4.2 Results
11.5 Indexing Iris Images
11.5.1 Indexing of Iris Database Based on Local Features
11.5.2 Results
11.6 Indexing Signature Images
11.6.1 KD-Tree-Based Signature Database Indexing
11.6.2 Results
11.7 Conclusion
References
Chapter 12 Iris Segmentation in theΒ Wild Using Encoder-Decoder-Based Deep Learning Techniques
12.1 Introduction
12.2 Deep Learning for Segmentation
12.3 Related Work
12.3.1 Non-Deep Learning-Based Methodologies
12.3.2 Deep Learning-Based Methodologies
12.4 Data Sets and Evaluation Metrics
12.4.1 Data sets
12.4.2 CASIA
12.4.2.1 UBIris v1 and UBIris v2
12.4.2.2 NICE-I and NICE-II
12.4.2.3 ND-Iris-0405
12.4.2.4 IITD
12.4.2.5 CSIP
12.4.2.6 MICHE-I and MICHE-II
12.4.2.7 SBVPI
12.4.2.8 IRISSEG-CC
12.4.2.9 IRISSEG-EP
12.4.2.10 MMU1 and MMU2
12.4.2.11 OpenEDS
12.4.2.12 iBUG
12.4.3 Performance Metrics
12.4.3.1 Jaccard Index (JI)
12.4.3.2 Mean Segmentation Error
12.4.3.3 Nice2 Error
12.5 Experimentation
12.6 Challenges Identified and Further Direction
12.7 Conclusion
Acknowledgements
References
Chapter 13 PPG-Based Biometric Recognition: Opportunities with MachineΒ and Deep Learning
13.1 Introduction
13.2 Photoplethysmogram (PPG)
13.3 Literature Review
13.4 Multi-Feature Approach for PPG Biometric
13.4.1 Signal Acquisition
13.4.2 Baseline Wander and Noise Removal
13.4.3 Feature Extraction
13.4.3.1 Pulse Extraction and Normalisation
13.4.3.2 First- and Second-Order Derivatives
13.4.3.3 Autocorrelation
13.5 Classification
13.6 Experiments and Results
13.7 Conclusions
References
Chapter 14 Current Trends of Machine Learning Techniques in Biometrics and its Applications
14.1 Introduction
14.1.1 Biometric Systems
14.1.2 Brain Stroke
14.1.2.1 Risk Factors
14.1.2.2 Blood Pressure
14.1.2.3 Heart Disease
14.1.2.4 Diabetes Mellitus
14.1.2.5 Cholesterol
14.1.2.6 Smoking
14.1.2.7 Alcohol
14.1.2.8 Other Risk Factors
14.1.3 Face Recognition
14.1.4 Motivation to Machine Learning Techniques
14.2 Related Work
14.2.1 Review on Brain Stroke
14.2.2 Review on Face Recognition
14.2.3 Brain Stroke Prediction System
14.2.3.1 Image Acquisition
14.2.3.2 Pre-processing
14.2.3.3 Feature Extraction
14.2.3.4 Classification Using Machine Leaning Algorithms
14.2.3.5 Construction of Convolutional Neural Network
14.2.4 Face-Recognition System
14.3 Discussion and Results
14.3.1 Performance of Brain Stroke
14.3.2 Performance of Face Recognition
14.4 Future Scope
14.5 Conclusion
References
Index