A Secure Template Generation Scheme for Palmprint Recognition Systems

Abstract

With the development of more and more systems which provide service based on the identity of a person, the importance of personal identification is growing. Providing authorized users with secure access to the services is a challenge to the personal identification systems. There are several conventional means for personal identi¯cation which include passports, keys, tokens, access cards, personal identification number (PIN), passwords. Unfortunately, passports, keys, access cards, tokens, can be lost, stolen or duplicated, and passwords, PINs can be forgotten,cracked or shared. These drawbacks cause a great loss to the concerned. Biometric systems are proving to be an e±cient solution to this problem. A biometric identity verification system tries to verify user identities by comparing some sort of behavioral or physiological trait of the user to a previously stored sample of the trait. The recent developments in the biometrics area have lead to smaller, faster and cheaper systems, which in turn has increased the number of possible application areas for biometric identity verification. Palmprint can be one of the biometrics, used for personal identification or verification. As a small central part of the palmprint image is used for this purpose, so it is important to find that region of interest. We propose a rotation and translation invariant preprocessing scheme which finds the central part of the palmprints. As biometric systems are vulnerable to replay, database and brute-force attacks, such potential attacks must be analyzed before they are massively deployed in security systems. Along with security, also the privacy of the users is an important factor as the constructions of lines in palmprints contain personal characteristics. We propose a cryptographic approach to encrypt the palmprint images by an advanced Hill cipher technique for hiding the information on the palmprints. It also provides security to the palmprint images from being attacked by above mentioned attacks. So, during the template generation, the encrypted palmprint sub-images are first decrypted and then the features are extracted.