Beyond Frontal Face Recognition

Face recognition is one of the most researched subjects in computer vision. The attention it receives due to complexity problem. models have deal with a wide variety intraclass variations such as pose variations, facial expressions, effect aging, and natural occlusion different illumination. This challenge often referred pose-illumination-expression. Our brain performs face efficiently because we process holistically. Since achieving human-level accuracy ultimate goal, should ascertain whether computational model mimicking this approach would better tackle In research, developed learning that closely mimics way human visual cortex recognition. It decomposes task into two specialized sub-tasks. A generator holistic step, followed by classifier for together "holistic model." To problem, introduced use calculated distance features known configural information, which correlate frontal profile face. We compared against baseline current state-of-the-art (or classical models). experimental results show outperforming Multi-PIE dataset, 2.14% performed expected Labeled Wild dataset. ability recognize any orientation high will tremendous impact on biometric authentication liveness detection.