Kelvin S. Bryant

Genetic-Based Type II Feature Extraction for Periocular Biometric Recognition: Less is More

Given an image from a biometric sensor, it is important for the feature extraction module to extract an original set of features that can be used for identity recognition. This form of feature extraction has been referred to as Type I feature extraction. For some biometric systems, Type I feature extraction is used exclusively. However, a second form of feature extraction does exist and is concerned with optimizing/minimizing the original feature set given by a Type I feature extraction method. This second form of feature extraction has been referred to as Type II feature extraction (feature selection). In this paper, we present a genetic-based Type II feature extraction system, referred to as GEFE (Genetic & Evolutionary Feature Extraction), for optimizing the feature sets returned by Loocal Binary Pattern Type I feature extraction for periocular biometric recognition. Our results show that not only does GEFE dramatically reduce the number of features needed but the evolved features sets also have higher recognition rates.

Minimizing the number of bits needed for iris recognition via Bit Inconsistency and GRIT

In this paper, we demonstrate how the concepts of Bit Inconsistency and Genetic Search can be used to minimize the number of iris code bits needed for iris recognition. In addition, we compare two systems: GRIT-I (Genetically Refined Iris Templates I) and GRIT-II. Our results show that GRIT-I (by evolving the bit mask of iris templates) was able to reduce the number of iris code bits needed by approximately 30% on average. GRIT-II by contrast optimizes the bit mask as well as the iris code bits that have 100% consistency and 100% coverage with respect to the training set. GRIT-II was able to reduce the number of iris code bits needed by approximately 89%.

Genetic based LBP feature extraction and selection for facial recognition

This paper presents a novel approach to LBP feature extraction. Unlike other LBP feature extraction methods, we evolve the number, position, and the size of the areas of feature extraction. The approach described in this paper also attempts to minimize the number of areas as well as the size in an effort to reduce the total number of features needed for LBP-based face recognition. In addition to reducing the number of features by 63%, our approach also increases recognition accuracy from an average of 99.04% to 99.84%.

GEFE: genetic & evolutionary feature extraction for periocular-based biometric recognition

Personal identification using an individuals periocular skin texture (e.g. the texture of the skin around the eye) is a promising and exciting new biometric modality [11]. For the application presented in this paper, local binary patterns (LBPs) are used to extract 1416 features from the periocular regions of images within the Face Recognition Grand Challenge (FRGC) dataset. GEFE (Genetic & Evolutionary Feature Extraction) is then used to evolve optimized subsets of the original feature set. Our results show that not only do the evolved subsets consist of approximately 50% fewer features but they also have higher recognition rates.

Genetic & Evolutionary Biometric Security: Disposable Feature Extractors for Mitigating Biometric Replay Attacks

Abstract Biometric-based access control systems (BACSs) are vulnerable to replay attacks. Replay attacks occur when a biometric template is intercepted and maliciously used to gain unauthorized access to a system. In this paper, we introduce a Genetic and Evolutionary Biometric Security (GEBS) application which uses a Genetic and Evolutionary Computation to develop disposable Feature Extractors (FEs) in an effort to mitigate replay attacks. We describe how a previously developed system known as GEFE (Genetic and Evolutionary Feature Extraction) can be used to evolve unique and disposable FEs for users of BACS. Furthermore, we propose two access control protocols based on the use of disposable FEs and/or their resulting templates (also referred to as feature vectors (FVs)). In our proposed protocols, FEs/FVs are used to authenticate the identity of individuals and are then discarded. Our results show that this GEBS application can be successfully used to mitigate biometric replay attacks.

Hybrid GAs for Eigen-based facial recognition

In this paper, we have performed an evaluation of genetic-based feature selection and weighting on the PCA-based face recognition. This work highlights the first attempt of applying Genetic Algorithm (GA) based feature selection on the Eigenface method. The results show that genetic-based feature selection reduces the number of features needed by approximately 50% while improving the identification accuracy over the baseline. Genetic-based feature weighting significantly improves the accuracy from an 87.14% to a 92.5% correct recognition rate.

A comparison of genetic feature selection and weighting techniques for multi-biometric recognition

In this paper, we compare genetic-based feature selection (GEFeS) and weighting (GEFeW) techniques for multi-biometric recognition using face and periocular biometric modalities. Our results show that fusing face and periocular features outperforms face-only and periocular-only biometric recognition. Of the two genetic-based approaches, GEFeW outperforms GEFeS.

GEC-based multi-biometric fusion

In this paper, we use Genetic and Evolutionary Computation (GEC) to optimize the weights assigned to the biometric modalities of a multi-biometric system for score-level fusion. Our results show that GEC-based multi-biometric fusion provides a significant improvement in the recognition accuracy over evenly fused biometric modalities, increasing the accuracy from 90.77% to 95.24%.

Genetic and evolutionary methods for biometric feature reduction

In this paper, we investigate the use of Genetic and Evolutionary Computations (GECs) for feature selection (GEFeS), weighting (GEFeW), and hybrid weighting/selection (GEFeWS) in an attempt to increase recognition accuracy as well as reduce the number of features needed for biometric recognition. These GEC-based methods were first applied to a subset of 105 subjects taken from the Facial Recognition Grand Challenge (FRGC) dataset (FRGC-105) where several feature masks were evolved. The resulting feature masks were then tested on a larger subset taken from the FRGC dataset (FRGC-209) in an effort to investigate how well they generalise to unseen subjects. The results suggest that our GEC-based methods are effective in increasing the recognition accuracy and reducing the features needed for recognition. In addition, the evolved FRGC-105 feature masks generalised well on the unseen subjects within the FRGC-209 dataset.

Genetic & Evolutionary Type II feature extraction for periocular-based biometric recognition

One of the most important modules of any bio-metric system is the feature extraction module. Given a sample it is important for the feature extraction method to extract a rich set of features that can be used for identity recognition. This form of feature extraction has been referred to as Type I feature extraction and for some biometric systems it is used exclusively. However, a second form of feature extraction does exist and is concerned with optimizing/minimizing the original feature set given by a Type I feature extraction method. This second form of feature extraction has been referred to as Type II feature extraction (also known as feature selection). In this paper, we compare two GEC-based Type II feature extraction methods as applied to periocular-based recognition, an exciting new area of research within the Biometric research community that to date has used Type I feature extraction exclusively. Our results show that GEC-based Type II feature extraction is effective in optimizing recognition accuracy as well as minimizing the overall feature set size.