Although automatic faces recognition has shown success for high-quality images under controlled conditions, for video-based\r\nrecognition it is hard to attain similar levels of performance.We describe in this paper recent advances in a project being undertaken\r\nto trial and develop advanced surveillance systems for public safety. In this paper, we propose a local facial feature based framework\r\nfor both still image and video-based face recognition. The evaluation is performed on a still image dataset LFW and a video\r\nsequence dataset MOBIO to compare 4 methods for operation on feature: feature averaging (Avg-Feature), Mutual Subspace\r\nMethod (MSM), Manifold to Manifold Distance (MMS), and Affine Hull Method (AHM), and 4 methods for operation on\r\ndistance on 3 different features. The experimental results show thatMulti-region Histogram (MRH) feature ismore discriminative\r\nfor face recognition compared to Local Binary Patterns (LBP) and raw pixel intensity. Under the limitation on a small number of\r\nimages available per person, feature averaging ismore reliable thanMSM,MMD, andAHM and ismuch faster. Thus, our proposed\r\nframeworkââ?¬â?averaging MRH feature is more suitable for CCTV surveillance systems with constraints on the number of images\r\nand the speed of processing.
Loading....