Kushan Ahmadian

Dealing with biometric multi-dimensionality through chaotic neural network methodology

Acquiring a group of different biometrics characteristic and specifications results in a number of issues that should be addressed in a modern biometric system. One of the common problems is the high dimensionality of the data, which may impact negatively the biometric system performance. The complexity of data is rarely considered in multimodal biometric systems due to the gap between recently developed dimensionality reduction techniques in data mining and data analysis of biometric features. To remedy the situation, this paper proposes a unique methodology for shrinking down the finite search space of all possible subspaces. The approach also utilises the function approximation capabilities of chaotic neural networks to act as an associative memory to learn the biometric patterns. In summary, the contribution of this paper is in novel methodology based on the axis-parallel dimension reduction technique and chaotic neural network to improve the performance and circumvention of biometric system.

On-demand chaotic neural network for broadcast scheduling problem

This paper presents a novel approach to optimizing network packet transfer scheme through introducing a new method for on-demand chaotic noise injection strategy for the Broadcast Scheduling Problem (BSP). Packet radio networks have many applications, while finding an optimized scheduling to transmit data is proven to be a NP-hard problem. The objective of the proposed method is to find an optimal time division multiple access (TDMA) frame, based on maximizing the channel utilization. The proposed method benefits from an on-demand noise injection policy, which injects noise based on the status of neuron and its neighborhoods. The method is superior to other Noise Chaotic Neural Networks (NCNN) that suffer from blind injection policy. The experimental result shows that, in most cases, the proposed on-demand noise injection algorithm finds the best solution with minimal average time delay and maximum channel utilization.

A multi-modal approach for high-dimensional feature recognition

Over the past few decades, biometric recognition firmly established itself as one of the areas of tremendous potential to make scientific discovery and to advance state-of-the- art research in security domain. Hardly, there is a single area of IT left untouched by increased vulnerabilities, on-line scams, e-fraud, illegal activities, and event pranks in virtual worlds. In parallel with biometric development, which went from focus on single biometric recognition methods to multi-modal information fusion, another rising area of research is virtual world’s security and avatar recognition. This article explores links between multi-biometric system architecture and virtual worlds face recognition, and proposes methodology which can be of benefit for both applications.

A Novel Multi-modal Biometric Architecture for High-Dimensional Features

Dealing with high-dimensional data has an important role in a number of areas, including biometric recognition in both real world and emerging virtual reality applications. Acquiring a group of different biometrics with various characteristics and specifications results in a number of issues that should be addressed, while developing such multi-modal recognition system. In this paper, we propose a novel Multi-Modal Biometric System based on neural network paradigm which utilizes the ear and face features and has unique method to train different classifiers based on each feature set. The aggregation result depicts the final decision over the recognized identity. In order to train accurate set of classifiers, the subspace clustering method has been used to overcome the problem of high dimensionality of the feature space. The proposed system is based on a new methodology for shrinking down the finite search space of all possible subspaces by focusing on axis-parallel subspaces which is a novel approach in data clustering for biometric dataset. The experimental results over the FERET dataset show the superiority of the proposed method over several dimensionality reduction methods.

Chaotic neural network for biometric pattern recognition

Biometric pattern recognition emerged as one of the predominant research directions inmodern security systems. It plays a crucial role in authentication of both real-world and virtual reality entities to allow system to make an informed decision on granting access privileges or providing specialized services. The major issues tackled by the researchers are arising from the ever-growing demands on precision and performance of security systems and at the same time increasing complexity of data and/or behavioral patterns to be recognized. In this paper, we propose to deal with both issues by introducing the new approach to biometric pattern recognition, based on chaotic neural network (CNN). The proposed method allows learning the complex data patterns easily while concentrating on the most important for correct authentication features and employs a unique method to train different classifiers based on each feature set. The aggregation result depicts the final decision over the recognized identity. In order to train accurate set of classifiers, the subspace clustering method has been used to overcome the problem of high dimensionality of the feature space. The experimental results show the superior performance of the proposed method.

On-Demand Chaotic Neural Network for Broadcast Scheduling Problem

An on-demand chaotic noise injection strategy for Broadcast Scheduling Problem (BSP) based on an adjacency matrix is described in this paper. Packet radio networks have many applications especially for military purposes while finding an optimized scheduling to transmit data is proven to be a NP-hard domain problem. The objective of the proposed method is to find an optimal time division multiple access (TDMA) frame based on maximizing channel utilization. The proposed method benefits from an on-demand noise injection policy which, unlike previous Noise Chaotic Neural Networks (NCNN) that suffers from blind injection policy, injects noise based on the status of neuron and its neighborhoods. The experimental result shows that in most cases the on-demand noise injection finds the best solution with minimal average time delays and maximum channel utilization in comparison to previous methods.

Multi-objective Evolutionary Approach for Biometric Fusion

In recent years, a noticeable amount of research has been focused on biometric fusion. A new area is looking at utilization of AdaBoost-type learning methods in biometric fusion domain. These methods rely on an idea that by selecting a variety of biometric classifiers the error rate can be reduced. This paper presents a new evolutionary algorithm based on the multi-objective genetic approach, which automatically preserves diversity in face detection system. The proposed algorithm creates classifiers based on the amount of error computed for each class, and then uses multi-objective genetic algorithm to combine them to produce a set of powerful ensembles. The application is developed for face detection biometric system.

Using Duality and Hopfield Neural Network for Delaunay Triangulation Based Fingerprint Matching

Abstract: In this paper we present a new method for fingerprint matching which is based on the calculation of Delaunay Triangulation (DT) of the minutiae set. The obtained DT is transformed to a set of points in the discretized space using duality. This translation results in a sampling method be acquiring which the system tolerates displacement and noise of the input image. Finally a Hopfield Neural Network (HNN) is used to learn the obtained pattern. Experimental results show a significant improvement in the False Rejection Rate over both the traditional DT-based approach and the direct HNN application.

Axis-Parallel dimension reduction for biometric research

The objective of this research is to present a novel methodology based on the axis-parallel dimension reduction technique and chaotic neural network to improve the performance and circumvention of multi-modal biometric system. The proposed methodology for dimensionality-reduction and chaotic neural network learner on example of face, ear and fingerprint biometric was presented in the methodology section. This paper validates the proposed methodology by providing experimentation results. First subsection showcases advantages of chaotic neural network for fingerprint recognition (accuracy and circumvention). Next subsection compares results of the proposed multi-biometric system based on axis-parallel dimensionality reduction. The experiments demonstrate that the proposed dimensionality reduction and associative memory training methodology outperforms other commonly used techniques in both FAR (False Accept Rate) and FRR (False Reject Rate) both individually and if used together. Finally, the last section proposes the alternative multimodal system architecture with additional features that can further improve system performance in terms of accuracy and circumvention.

Chaotic Neural Network and Multidimensional Data Analysis in Biometric Applications

In this book chapter, a novel biometric system from the normalisation level up to the verification level is developed, tested and verified against other multimodal and unimodal systems. The main advantage of a new architecture is in flexibility of combining various features from multimodal biometrics in a new way, suitable for neural-network learner. The system utilises associative memories and pattern matchers as learners of biometric data, but the main advantage of a new architecture is increased resistance to noise and ability of system to compensate for an absence of some biometric traits. Detailed experimental analysis of pros and cons of such system is also provided.