Ying Han Pang

An efficient method for human face recognition using wavelet transform and Zernike moments

This paper presents a method of combining wavelet transforms (WT) and Zernike moments (ZM) as a feature vector for face recognition. Wavelet transform, with its approximate decomposition is used to reduce the noise and produce a representation in the low frequency domain, and hence making the facial images insensitive to facial expression and small occlusion. The Zernike moments, on the other hand, is selected as feature extractor due to its robustness to image noise, geometrical invariants property and orthogonal property. The simulation results on Essex database indicates that higher order degree of WT combine with ZM achieve better performance with respect to recognition rate rather than using WT or ZM alone. The optimum result is obtained for ZM of order 10 with Daubechies orthonormal wavelet filter of order 7 in the first decomposition level. It can achieve the verification of 94.26%.

Touch-less Fingerprint Recognition System

Touch-less fingerprint recognition is regarded as a viable alternative to contact-based fingerprint recognition technology. It provides a near ideal solution to the problems in terms of hygienic, maintenance and latent fingerprints. In this paper, we present a touch-less fingerprint recognition system by using a digital camera. Specifically, we address the constraints of the fingerprint images that were acquired with digital camera, such as the low contrast between the ridges and the valleys in fingerprint images, defocus and motion blurriness. The system comprises of preprocessing, feature extraction and matching stages. The proposed preprocessing stage shows the promising results in terms of segmentation, enhancement and core point detection. Feature extraction is done by Gabor filter and the favorable verification results are attained with the Support Vector Machine.

Palmprint authentication with Zernike moment invariants

Nowadays, palmprint verification is a novel and one of the most reliable biometrically based technology in applications of personal identification and authentication due to its high stability and uniqueness. The principal features of both Chinese character and palmprint are based on the line structure as feature descriptor. This idea prompts us to implement one of the most commonly used Chinese character recognition and reconstruction feature extraction technique, namely Zernike moment invariants (ZMI), in the application of human palmprint authentication. This technique is able to define the statistical and geometrical features containing the line structural information about palmprint. An experimental study about the verification rate of the palmprint authentication system using the Zernike moment invariants has been discussed here. Zernike moment invariants orthogonality and translation, rotation and scale invariant properties promote itself as a widely used feature extraction alternative in a broad spectrum of applications in image analysis.

Digital camera based fingerprint recognition

Touch-less fingerprint recognition deserves increasing attention as it lets off the problems of deformation, maintenance, latent fingerprint problems and so on that still exist in the touch-based fingerprint technology. However, problems such as the low ridges-valleys contrast in the fingerprint images, defocus and motion blurriness raise when developing a digital camera based fingerprint recognition system. The system comprises of preprocessing, feature extraction and matching stages. The proposed preprocessing stage presents the promising results in terms of segmentation, enhancement and core point detection. Feature extraction is done by Gabor filter followed by principle component analysis (PCA) and the favorable verification results are attained with Cosine Angle.

Supervised Locally Linear Embedding in face recognition

Locally Linear Embedding (LLE), which has recently emerged as a powerful face feature descriptor, suffers from a limitation. That is class-specific information of data is lacked of during face analysis. Thus, we propose a supervised LLE technique, known as class-label Locally Linear Embedding (cLLE), to overcome the problem. cLLE is able to discover the nonlinearity of high-dimensional face data by minimizing the global reconstruction error of the set of all local neighbors in the data set. cLLE utilizes user class-specific information in neighborhoods selection and thus preserves the local neighborhoods. Since the locality preservation is correlated to the class discrimination, the proposed cLLE is expected superior to LLE in face recognition. Experimental results on three face databases: ORL, AR and Yale databases, demonstrate that the proposed technique obtains better recognition performance than PCA and LLE.

Enhanced pseudo Zernike moments in face recognition

This paper presents an approach to boost the performance of pseudo Zernike moments in face recognition. This approach is a hybrid of a kernel trick, discriminant function and pseudo Zernike moments (PZM), namely as Kernel-based Fisher Pseudo Zernike Moments (KFPZM). KFPZM maps the moment-based features into a high dimensional feature space via kernel function for disclosing the underlying variables which carry significant information about the image. Then, it performs discriminant analysis onto the mapped features to enhance the discrimination power via Fishers Linear Discriminant (FLD). Experimental results show that the proposed method outperforms the sole PZM and the integrated FLD with PZM methods, achieving recognition rate of 98.11% and 93.03% in the face databases with facial expression variations and illumination variations, respectively.

Personal authenticator on the basis of two-factors: palmprint features and tokenized random data

This paper presents a novel two-factor authenticator which hashes tokenized random data and moment based palmprint features to produce a set of private binary string, coined as Discrete-Hashing code This novel technique requires two factors (random number + authorized biometrics) credentials in order to access the authentication system Absence of either factor will just handicap the progress of authentication Besides that, Discrete-Hashing also possesses high discriminatory power, with highly correlated bit strings for intra-class data Experimental results show that this two-factor authenticator surpasses the classic biometric authenticator in terms of verification rate Our proposed approach provides a clear separation between genuine and imposter population distributions This implies that Discrete-Hashing technique allows achievement of zero False Accept Rate (FAR) without jeopardizing the False Reject Rate (FRR) performance, which is hardly possible to conventional biometric systems.

Binarized revocable biometrics in face recognition

This paper proposes a novel revocable two-factor authentication approach which combines user-specific tokenized pseudo-random bit sequence with biometrics data via a logic operation. Through the process, a distinct binary code per person, coined as bio-Bit, is formed. There is no deterministic way to acquire bio-Bit without having both factors. This feature offers an extra protection layer against biometrics fabrication since bio-Bit authenticator is replaceable via token replacement. The proposed method also presents functional advantages of obtaining zero equal error rate and yielding a clean separation between the genuine and imposter populations. Thereby, false accept rate can be eradicated without suffering from the increased occurrence of false reject rate.

Two-factor cancelable biometrics authenticator

Biometrics-based authentication system offers advantages of providing high reliability and accuracy. However, the contemporary authentication system is impuissance to compromise. If a biometrics data is compromised, it cannot be replaced and rendered unusable. In this paper, a cancelable biometrics-based authenticator is proposed to solve this irrevocability issue. The proposed approach is a two-factor authentication system, which requires both of the random data and facial feature in order to access the system. In this system, tokenized pseudo-random data is coupled with moment-based facial feature via inner product algorithm. The output of the product is then discretized to generate a set of private binary code, coined as 2factor-Hashing code, which is acted as verification key. If this biometrics-based verification key is compromised, a new one can be issued by replacing a different set of random number via token replacement. Then, the compromised one is rendered completely useless. This feature offers an extra protection layer against biometrics fabrication since the verification code is replaceable. Experimental results demonstrate that the proposed system provides zero Equal Error Rate in which there is a clear separation in between the genuine and the imposter distribution populations.

In-air hand gesture signature recognition system based on 3-dimensional imagery

A traditional online handwritten signature recognition system requires direct contact to acquisition device and usually will leave a traceable print on the surface. This made a signature possible and vulnerable to certain attempts of tracking and imitated. Looking into this shortfall, this paper proposes a novel approach to recognise an individual based on his/ her in-air hand motion while signing his/her signature. In this study, a low-cost acquisition device – Microsoft Kinect sensor is adopted to capture an image sequence of hand gesture signature. Palm region is first located and segmented through a predictive palm segmentation algorithm, which are then combined to generate a volume data. The volume data is condensed and reduced into a motion representation image by means of Motion History Image (MHI), which produces rich motion and temporal information. Several features are extracted from the MHI for empirical evaluation. Two classical recognition modes – identification and verification, are testified with an in-house database (HGS database). The proposed system achieves 90.4% identification accuracy and 3.22% equal error rate in verification mode. The experimental results substantiated the potential of the proposed system.