Research On Face Detection Method

Images containing faces are essential to intelligent vision-based human computer interaction, and research efforts in face processing include face recognition, face tracking, pose estimation, and expression recognition. However, all of these researching directions involve in one problem-face detection and location. In other words, before this face processing, we must know faces' locations and scales. Consequently, to build an automated face processing system which analyzes the information contained in face images, robust and efficient face detection algorithms are required. The research on face detection has lasted for more than twenty years. But, up to now, due to the complexity of the purpose such as the diversity of face patterns, variable lighting condition and so on, many researches can not resolve the problem completely even if they have studied it for long time, hi this thesis, the author has done some work on the face detection.The work includes:(1) The improvement of AdaBoost learning algorithmUsing a training set of positive and negative image, Paul Viola has used AdaBoost to build a robust scale-invariant image classifier for face detection. Based on the deeply analysis of AdaBoost learning, this paper improves the classical training methods in two ways: Firstly, the basic and over-complete set of feature is extended by an set of rotated features ,which add additional domain-knowledge to the learning framework and which is otherwise hard to learn. Secondly, this paper derives a new method to optimize the learning procedure, which compute the features only twice and use more less space in memory.(2)Face detection based on skin colorAmong many color spaces, this paper used YCbCr components. Since in the YCbCr color space, the luminance information is contained in Y component; and the chrominance information is in Cb and Cr. Therefore, the luminance information can be easily de-embedded. This paper proposes a face detection algorithm for color images in the presence of varying lighting conditions as well complex backgrounds. Based on a novel lighting compensation, our method detects skin regions over the entire image, and then generates face candidates based on the spatial arrangement of these skin patches. The algorithm constructs eye and mouth maps for verifying each face candidate. Experimental results demonstrate successful face detection over a range of facial variations in color, position, scale and expression in color images.(3) A face detector based on skin color segmenting pre-processing and multiple cascade of classifiers based on AdaBoost validation.Multiple cascades of classifiers based on AdaBoost statistical learning method have been applied successfully in face detection problem. The advantage of using multiple cascades for face detection is the feasibility of training a system to capture the complex class conditional density of face patterns. However, one drawback is that it need apply an exhaustive window scanning technique to input image for possible face locations at all scales. This reduces the executing efficiency of the detector in great extent and restricts it to apply to some real-time face detection application. The results of experiments show the new face detector achieves a higher executing efficiency and fewer false detect than traditional multiple cascades detector.