Research On Algorithm For Human Face Detection In Colorful Image With Complex Background

Recognition and analysis of human faces can be widely used in such fields as personal identification recognition, safety inspection, human-computer interaction, expression analysis, lip-reading, etc. As the basis, necessity and prerequisite of human face recognition and analysis, human face detection with the help of computer attracted people's attention at a very early time. Along with the generalization of computer application, the improvement of computer performance and achievements in image processing and pattern recognition fields, human-face concerned applications is getting nearer and nearer to practical use, thus resulting in more and more emphasis pver research for human face detection.There have already been a lot of successful algorithms for human-face detection in simple background. However, as to images in complex background, existing algorithms either are slow at detecting or easily lead to detection failure and detection fault. In this paper, first an overview and analysis is given of existing typical algorithms for human detection, then a multi-stage algorithm is proposed to detect human faces whose numbers, sizes and positions are all unknown. The algorithm uses . complexion model to preliminarily extract human-face regions. Then use adaptive region-based complexion re-extracting algorithm to extract human-faces again, afterthat, a region shape is used to filter the faces acquired from above two steps in order to identify the right people and finally the eyes are detected for further verifying the detecting result. Since the application of complexion model help reduce the regions likely to be selected in the complex background and the subsequent algorithm is simple and useful, it shows a higher speed for detection than edge-based detection and template matching algorithm. Adaptive region-based complexion re-extraction algorithm reduces the sensitivity of complexion model to the light conditions, thus ameliorating the common deficiencies existing in all complexion-model-based detection algorithms, namely light-sensitive, so as to reduce the rate of detection failure.In this paper, simple and useful algorithms in all phases are given, which is well worth referencing by other similar applications.