22 Multi-Modal Human Verification Using Face and Speech Changhan Park 1 and Joonki Paik 2 1 Advanced Technology R&D Center, Samsung Thales Co., Ltd., 2 Graduate School of Advanced Imaging Science, Multimedia, and Film Chung-Ang University, Seoul Korea 1. Introduction Human biometric characteristics are unique, so it can hardly be duplicated (Kong et al. 2005). Such information includes; facial, speech, hands, body, fingerprints, and gesture to name a few. Face detection and recognition techniques are proven to be more popular than other biometric features based on efficiency and convenience (Kriegman et al. 2002; Liu et al. 2002). It can also use a low-cost personal computer (PC) camera instead of expensive equipments, and require minimal user interface. Face authentication has become a potential a research field related to face recognition. Face recognition differs from face authentication because the former has to determine the identity of an object, while the latter needs to verify the claimed identity of a user. Speech (Gu and Thomas 1999) is one of the basic communications, which is better than other methods in the sense of efficiency and convenience. Each a single biometric information, however, has its own limitation. For this reason, we present a multimodal biometric verification method to reduce false acceptance rate (FAR) and false rejection rate (FRR) in real-time. There have been many approaches for extracting meaningful features. Those include principal component analysis (PCA) (Rowley et al. 1998), neural networks (NN) (Rowley et al. 1998), support vector machines (SVM) (Osuna et al. 1997), hidden markov models (HMM) (Samaria and Young 1994), and linear discriminant analysis (LDA) (Belhumeur et al. 1997). In this chapter, we use the PCA algorithm with unsupervised learning to extract the face feature. We also use the HMM algorithm for extracting speech feature with supervised learning. This chapter is organized as follows: Section 2 and 3 describe feature extraction of face and speech using the PCA and HMM algorithms, respectively. Section 4 presents the design and structure of the proposed system. Section 5 presents experimental, and Section 6 concludes the paper with future research topics. 2. Face Extraction and Recognition In this section, the proposed face extraction and recognition method will be presented. The proposed method can deal with both gray and color images. Depending on the type of images, an additional preprocessing step may be included so that facial features can be detected more easily. Source: Face Recognition, Book edited by: Kresimir Delac and Mislav Grgic, ISBN 978-3-902613-03-5, pp.558, I-Tech, Vienna, Austria, June 2007 Open Access Database www.i-techonline.com