Block Statistical Features-based Face Verification on Second Generation Identity Card

Face verification has been extensively studied in recent decades. Nevertheless, face verification for identity card has subject to relatively little attention. This paper proposes a block statistical features(BSF) learning method combining with local Gabor binary pattern(LGBP) for face verification in both second generation identity card(2nd ID card) and video set, which show many differences in biometric caused by age gap and image acquisition conditions. To alleviate computation complexity of Gabor transformation, we exploit energy check Gabor filters to speed up calculation. Specially, the verification rate of our approach on NEU-ID database achieves 97.71 %. It has a comparable performance with lower computation complexity.