An efficient computational intelligence technique for affine-transformation-invariant image face detection, tracking, and recognition in a video stream

Abstract

While there are many current approaches to solving the difficulties that come with detecting, tracking, and recognizing a given face in a video sequence, the difficulties arising when there are differences in pose, facial expression, orientation, lighting, scaling, and location remain an open research problem. In this paper we present and perform the study and analysis of a computationally efficient approach for each of the three processes, namely a given template face detection, tracking, and recognition. The proposed algorithms are faster relatively to other existing iterative methods. In particular, we show that unlike such iterative methods, the proposed method does not estimate a given face rotation angle or scaling factor by looking into all possible face rotations or scaling factors. The proposed method looks into segmenting and aligning the distance between two eyes' pupils in a given face image with the image x-axis. Reference face images in a given database are normalized with respect to translation, rotation, and scaling. We show here how the proposed method to estimate a given face image template rotation and scaling factor leads to real-time template image rotation and scaling corrections. This allows the recognition algorithm to be less computationally complex than iterative methods.