FACE DETECTION AND RECOGNITION FROM VIDEOS USING CASCADED DEEP LEARNING AND BAYESIAN LEARNING TECHNIQUE

Now a days, security based applications are developed widely and these systems are adopted in various real-time applications. Visual surveillance is considered as a most promising technique where certain objects can be detected, tracked and recognized using computer vision based approaches. In this field, face detection and recognition is considered as the important part of surveillance system. Several approaches have been developed for face recognition but existing approaches are applied on the face data. Recently, video face detection techniques are also introduced which provides more information to improve the security system. Deep learning achieves substantial improvements in face detection. However, the existing methods need to input fixed-size images for image processing and most methods use a single network for feature extraction, which makes the model generalization ability weak. In response to the above problems, our framework leverages a cascaded architecture with three stages of deep convolutional networks to improve detection performance. The network can predict face in a coarse-to-fine manner. The proposed method takes the Three-Patch Local Binary Pattern (TPLBP) texture feature which has excellent performance in face analysis as the input of the network. The learning process is developed using Bayesian learning approach is developed. The proposed approach is implemented on benchmark datasets such as IARPA Janus Benchmark A (IJB-A), the YouTube Face dataset and the Celebrity-1000 dataset. A comparative performance is carried out which shows the robust performance of proposed approach.