Three measures for secure palmprint identification

⎯ Most previous research in the area of personal authentication using the palmprint as a biometric trait has concentrated on enhancing accuracy yet resistance to attacks is also a centrally important feature of any biometric security system. In this paper, we address three relevant security issues: template re-issuances, also called 1 cancellable biometrics, replay attacks, and database attacks. We propose to use a random orientation filter bank (ROFB) as a feature extractor to generate noise-like feature codes, called Competitive Codes for templates re-issuances. Secret messages are hidden in templates to prevent replay and database attacks. This technique can be regarded as template watermarking. A series of analyses is provided to evaluate the security levels of the measures.