Abstract—Electronic commerce is growing rapidly with on-line sales already heading for hundreds of billion dollars per year. Due to the huge amount of money transferred everyday, an increased security level is required. In this work we present the architecture of an intelligent speaker verification system, which is able to accurately verify the registered users of an e-commerce service using only their voices as an input. According to the proposed architecture, a transaction-based e-commerce application should be complemented by a biometric server where customer’s unique set of speech models (voiceprint) is stored. The verification procedure requests from the user to pronounce a personalized sequence of digits and after capturing speech and extracting voice features at the client side are sent back to the biometric server. The biometric server uses pattern recognition to decide whether the received features match the stored voiceprint of the customer who claims to be, and accordingly grants verification. The proposed architecture can provide e-commerce applications with a higher degree of certainty regarding the identity of a customer, and prevent impostors to execute fraudulent transactions. Keywords—Speaker Recognition, Biometrics, E-commerce security. I. INTRODUCTION RANSACTION-BASED Internet applications, such as e- commerce, e-business and e-finance are continuously growing and gaining a considerable part of the marketplace. Providing access to information has been handled so far by implementation of personal identification numbers (PINs), cards or tokens [1]. The problem with these is that they do not ensure high security in cases such as transactions encountered in e-commerce. To further increase e-commerce growth and public respect, higher security protection should be provided to customers. By integrating advanced biometric verification in an Internet application, secure, low-risk and convenient transactions can be executed. Biometric verification means the automated use of physiological or behavioral characteristics, such as iris, face, signature, finger, or voice, to verify one’s claimed identity [2]. Manuscript received May 31, 2006. This work was supported in part by the Cyprus Research Promotion Foundation and the General Secretariat for Research and Technology of the Ministry of Development of Greece under the Grant KY-EL/0603/77 and PLHRO/0603/01. A. Kounoudes is with ‘The Philips College’, Lamias 4-6 Nicosia, 2001 Cyprus (e-mail: tkounoudes@cytanet.com.cy). S. Mavromoustakos, is with ‘The Philips College’, Lamias 4-6 Nicosia, 2001 Cyprus (phone: +357.99.556939; e-mail: stephano@logosnet.cy.net.cy). Verification methods using biometrics can replace or complement conventional authorization mechanisms, namely passwords and personal identification numbers (PINs), for higher security applications. The main risk of traditional authorization methods is that passwords and PINs are sensitive to be stolen, guessed or retrieved by a person. Moreover, considering the amount of e-commerce services a customer uses that require a password, it is difficult for him/her to possess securely multiple and difficult to be guessed passwords. On the other hand, biometrics utilize intrinsic characteristics of a person and are not susceptible to fraud. Another advantage of biometric methods over the traditional ones is that the verification is not restricted to a binary decision, thus multiple levels of security can be posed. Comparing the biometric methods mentioned above, voice biometric systems, which are based on speaker verification, could be regarded as the most promising technique for being widely utilized in e-commerce applications. The increased presence of microphones devices and their low cost compared to other biometric acquisition devices makes voice biometrics the least expensive to deploy. Furthermore, voice biometrics do not invade customer’s privacy and users are more willing to cooperate (voice has not been used for individual tracking and monitoring). This paper, proposes an intelligent speaker verification process for enhancing security on e-commerce transactions using biometrics. The proposed method suggests the transmission of voice features instead of the whole speech signal to ensure maximum security and privacy and also save on bandwidth. The structure of this paper is as follows: Section II discusses biometric verification and its advantages and disadvantages. Section III provides an overview of the system. Section IV explains the proposed speaker verification process including the enrolment process and the verification process. This section also describes the tests performed to evaluate the performance of the system in real environment conditions. Finally, Section V sums up the findings of the paper and provides some concluding remarks. II. BIOMETRICS: ADVANTAGES AND DISADVANTAGES Biometrics is the science of using digital technology to identify the identity of individuals based on behavioral or physiological characteristics. By basing a security system on the physiological features rather than a few keystrokes or a password, the possibilities of fraud are drastically reduced. Intelligent Speaker Verification based Biometric System for Electronic Commerce Applications Anastasis Kounoudes, and Stephanos Mavromoustakos T World Academy of Science, Engineering and Technology International Journal of Economics and Management Engineering Vol:2, No:8, 2008 929 International Scholarly and Scientific Research & Innovation 2(8) 2008 scholar.waset.org/1307-6892/1289 International Science Index, Economics and Management Engineering Vol:2, No:8, 2008 waset.org/Publication/1289