Liveliness Detection in Real Time Videos using Color based Chromatic Moment Feature

摘要

Spoofing attack is one of the main security issues in Face Recognition System (FRS). There is a possibility of the FRS being spoofed by non-real faces such as printed photograph, replay video and photo displayed by a mobile device. The performance of the existing spoof detection is poor due to variation in illumination condition. This paper proposes an efficient Color based Chromatic Moment Features (CCMF) scheme for spoof detection. The reliability of liveliness is tested with different types of spoofing attacks like videos and recorded videos. The experimental results show that the proposed CCMF method gives better accuracy than the existing schemes.