Purdy Ho

Face recognition with support vector machines: global versus component-based approach

We present a component-based method and two global methods for face recognition and evaluate them with respect to robustness against pose changes. In the component system we first locate facial components, extract them and combine them into a single feature vector which is classified by a Support Vector Machine (SVM). The two global systems recognize faces by classifying a single feature vector consisting of the gray values of the whole face image. In the first global system we trained a single SVM classifier for each person in the database. The second system consists of sets of viewpoint-specific SVM classifiers and involves clustering during training. We performed extensive tests on a database which included faces rotated up to about 40/spl deg/ in depth. The component system clearly outperformed both global systems on all tests.

Face recognition: component-based versus global approaches

We present a component-based method and two global methods for face recognition and evaluate them with respect to robustness against pose changes. In the component system we first locate facial components, extract them, and combine them into a single feature vector which is classified by a support vector machine (SVM). The two global systems recognize faces by classifying a single feature vector consisting of the gray values of the whole face image. In the first global system we trained a single SVM classifier for each person in the database. The second system consists of sets of view-specific SVM classifiers and involves clustering during training. We performed extensive tests on a database which included faces rotated up to about 40° in depth. The component system clearly outperformed both global systems.

The Kullback-Leibler Kernel as a Framework for Discriminant and Localized Representations for Visual Recognition

The recognition accuracy of current discriminant architectures for visual recognition is hampered by the dependence on holistic image representations, where images are represented as vectors in a high-dimensional space. Such representations lead to complex classification problems due to the need to 1) restrict image resolution and 2) model complex manifolds due to variations in pose, lighting, and other imaging variables. Localized representations, where images are represented as bags of low-dimensional vectors, are significantly less affected by these problems but have traditionally been difficult to combine with discriminant classifiers such as the support vector machine (SVM). This limitation has recently been lifted by the introduction of probabilistic SVM kernels, such as the Kullback-Leibler (KL) kernel. In this work we investigate the advantages of using this kernel as a means to combine discriminant recognition with localized representations. We derive a taxonomy of kernels based on the combination of the KL-kernel with various probabilistic representation previously proposed in the recognition literature. Experimental evaluation shows that these kernels can significantly outperform traditional SVM solutions for recognition.

Biometric Authentication in Infrastructure Security

This paper will identify and recommend biometric technologies that provide strong authentication, convenient usability, and versatility, in order to meet the demand of enterprise infrastructure security systems. We aim at validating suitability for one or more mainstream applications. Finger scan technology is mature and widely available. The combination of a finger scan and smart card gives true dual-factor authentication that provides a greater degree of security and simplifies authentication for end users while preserving privacy. Speaker Verification is the most natural biometric technology to use with voice-based systems. It takes advantage of the ubiquitous voice-grade telephone channel. The combination of speaker verification and token technology can provide convenient and secure access to voice portal applications. We also discuss cultural, legal, and privacy issues based on religious objections, health concerns, legal restrictions, and regulations on the use of biometric technology.

A dual-factor authentication system featuring speaker verification and token technology

This paper presents a secure voice authentication system combining speaker verification and token technology. The dual-factor authentication system is especially designed to counteract imposture by pre-recorded speech and the text-to-speech voice cloning (TTSVC) technology, as well as to regulate the inconsistency of audio characteristics among different handsets. The token device generates and prompts a one-time passcode (OTP) to the user. The spoken OTP is then forwarded simultaneously to both a speaker verification module, which verifies the users voice, and a speech recognition module, which converts the spoken OTP to text and validates it. Thus, the OTP protects against recorded speech or voice cloning attacks and speaker verification protects against the use of a lost or stolen token device. We show the preliminary results of our Support Vector Machine (SVM)-based speaker verification algorithm, handset identification algorithm, and the system architecture of our design.