Andrew Teoh Beng Jin

Design and implementation of a contactless palm print and palm vein sensor

This paper presents an innovative contactless palm print and palm vein recognition system. We design a hand sensor that could capture the palm print and palm vein image using low-resolution web camera. Both the visible and infrared images can be captured at the same time, and we do not need specialized infrared sensor to image the vein pattern. The design of the device is simple and low-cost. The subject can be shielded completely from the complication of undergoing two separate acquiring processes. We allow subjects to position their hands freely above the sensor and they can move their hands during the acquisition process. In order to obtain clear image of the palm vascular pattern, we propose a novel image enhancement technique called local-ridge-enhancement (LRE). The proposed method removes illumination error while keeping good contrast between the print/vein pattern and the background image. Besides, we introduce a simple yet robust directional coding technique to encode the palm print and palm vein features in bit string representation. The bit string representation offers speedy template matching and enables more effective template storage and retrieval. The scores output by the palm print and palm vein experts are fused using Support Vector Machine. The fusion of these features yields promising result for practical implementation.

Fake-fingerprint detection using multiple static features

Recently, fake fingerprints have become a serious concern for the use of fingerprint recognition systems. We introduce a novel fake-fingerprint detection method that uses multiple static features. With regard to the usability of the method for field applications, we employ static features extracted from one image to determine the aliveness of fingerprints. We consider the power spectrum, histogram, directional contrast, ridge thickness, and ridge signal of each fingerprint image as representative static features. Each feature is analyzed with respect to the physiological and statistical distinctiveness of live and fake fingerprints. These features form a feature vector set and are fused at the feature level through a support vector machine classifier. For performance evaluation and comparison, a total of 7200 live images and 9000 fake images were collected using four sensors (three optical and one capacitive). Experimental results showed that proposed method achieved approximately 1.6% equal-error rate with optical-based sensors. In the case of the capacitive sensor, there was no test error when only one image was used for a decision. Based on these results, we conclude that the proposed method is a simple yet promising fake-fingerprint inspection technique in practice.

Fingerprint template protection with Minutia Vicinity Decomposition

Minutia vicinity representation was recently proposed by Yang & Busch to generate a protected fingerprint template scheme [14], the resultant protected template enjoys good accuracy and free from alignment. However, Yang and Busches scheme is highly likely reversible [18]. This paper proposed a new minutiae representation technique known as Minutia Vicinity Decomposition (MVD) whereby each minutia vicinity is decomposed into four minutia triplets. A set of geometrical invariant features can be extracted from the minutia triplet to construct a fingerprint template. The invariant features with random offsets salting mechanism enhance the reversibility, revocability as well as performance accuracy of the resultant protected fingerprint template. Promising experimental results on FVC2002 DB2 justify the feasibility of our proposed technique.

FuzzyHash: A Secure Biometric Template Protection Technique

Biometrics is likely to provide a new level of security to applications since application users will need to prove their identity when attempting access. Yet if the stored biometric template is compromised then the conventional biometric recognition system is vulnerable to privacy invasion, which is also a permanent loss because the biometric template is not replaceable. We introduce a new technique for protecting the biometric template called FuzzyHashing. It takes inspiration in the cryptographic hashing function, which endows it with properties that are desirable for improved security such as: one-way transformation, confusion, diffusion etc. After an enumeration of the design requirements of FuzzyHashing we demonstrate one of its realizations in the context efface biometrics.

Design and implementation of a contactless palm vein recognition system

This paper presents an innovative contactless palm vein recognition system. We design a hand sensor that could capture the palm vein image using low-resolution web camera. The design of the sensor is simple and low-cost, and we do not need to install specialized infrared sensor. We allow subjects to position their hands freely above the sensor and they can move their hands during the acquisition process. In order to obtain clear image of the palm vascular pattern, we propose a novel image processing technique called local-ridge-enhancement (LRE). The proposed method removes illumination error while keeping good contrast between the vein and the surrounding tissue. Apart from that, we present a new way to evaluate the image quality. Sometimes, the vein image does not appear clear due to the medical condition of the skin (like thick fatty tissue obstructing the subcutaneous blood vessels) and other environmental factor. These types of images are not suitable to be used for processing in our system. Therefore, we introduce the image quality checking procedure to evaluate the quality of the image before accepting it into our system. The proposed methodology improves the overall performance of the palm vein recognition system.

Locating geometrical descriptors for hand biometrics in a contactless environment

This paper proposes an innovative contactless hand geometry recognition system. We present a novel hand tracking approach to automatically detect and capture the geometrical features of the hand from low resolution video stream. No constraint is imposed and the subject can place his/her hand naturally on top of the sensor without touching any device. Conventional hand geometry systems require fairly precise positioning of the hand in order to obtain accurate measures of the hand. However, the proposed contactless approach does not fix any guidance pegs to help placing the hand at the right position when the image is acquired. As a result, the hand image may appear larger when the hand is placed near the sensor, and vice versa. Besides, the hand can be positioned at different angles. In other words, there is no way to obtain standard and constant hand measurements from this contactless setting. This research aims to deal with this complication when we have to get accurate measurements of the hand from images with varying sizes and directed at different orientations. Experiments show that our proposed method offers promising result for hand geometry recognition in a real-time contactless environment.

Probabilistic random projections and speaker verification

Biometrics is susceptible to non-revocable and privacy invasion problems. Multiple Random Projections (MRP) was introduced as one of the cancellable biometrics approaches in face recognition to tackle these issues. However, this technique is applicable only to 1D fixed length biometric feature vector but failed in varying size feature, such as speech biometrics. Besides, simple matching metric that used in MRP unable to offer a satisfactory verification performance. In this paper, we propose a variant of MRP, coined as Probabilistic Random Projections (PRP) in text-independent speaker verification. The PRP represents speech feature in 2D matrix format and speaker modeling is implemented through Gaussian Mixture Model. The formulation is experimented under two scenarios (legitimate and stolen token) using YOHO speech database. Besides that, desired properties such as one-way transformation and diversity are also examined.

Discriminative Discriminant Common Vector in face verification

Discriminant Common Vectors (DCV) is proposed to solve small sample size problem. Face recognition encounters this dilemma where number of training samples is always smaller than the data dimension. In literature, it is shown that DCV is efficient in face recognition. In this paper, DCV is enhanced for further boosting its discriminating power. This modified version is namely Discriminative Discriminant Common Vectors (DDCV). In this technique, a local Laplacian matrix of face data is computed. This matrix is used to derive a regularization model for computing discriminative class common vectors. Experimental results demonstrate that DDCV illustrates its effectiveness on face verification, especially on facial images with significant intra class variations.

Kernel-based Regularized Neighbourhood Preserving Embedding in face recognition

Face images always have significant intra-class variations due to different poses, illuminations and facial expressions. These variations trigger substantial deviation from the linearity assumption of data structure, which is essential in formulating linear dimension reduction technique. In this paper, we present a kernel based regularized graph embedding dimension reduction technique, known as kernel-based Regularized Neighbourhood Preserving Embedding (KRNPE) to address this problem. KRNPE first exploits kernel function to unfold the nonlinear intrinsic facial data structure. Neighbourhood Preserving Embedding, a graph embedding based linear dimension reduction technique, is then regulated based on Adaptive Locality Preserving Regulation Model, established in [7] to enhance the locality preserving capability of the projection features, leading to better discriminating capability and generalization performance. Experimental results on PIE and FERET face databases validate the effectiveness of KRNPE.

Secure Template Protection in Touch-less Based Fingeprint Verification System

Cancellable biometrics has been a challenging and essential approach to protect the privacy of biometric templates. Multiple Random Projections (MRP) is our formerly presented two-factor cancellable formulation. In that method, the biometric data is changed in a revocable but noninvertible manner by projecting every fixed length feature vector (extracted from the raw biometrics) onto a user-specific random subspace. In this paper, we propose a variant of MRP, namely Multiple Random Projections-Support Vector Machine (MRP-SVM). The MRPs template protection characteristics are inherited by MRP-SVM due to existence of the property of dot product and non-linear kernel. Furthermore, the verification performance is improved. This approach is verified using the touch-less based acquired fingerprint images. Touch-less based acquired images are free from latent fingerprint issues that can lead to fraudulent use. Hence, the security and privacy protection of fingerprint biometric templates is consolidated by the cancellable biometrics approach.