Face Detection Method Based On Low-Luminous Image

Face detection is a vital research area of pattern recognition and computer vision,and it has significant value of application.It always face to the image blurred and the illumination are not enough in the application of the face detection. Based on the deeply research at home and abroad of face detection and image enhancement,this paper research for the details enhancement of low illumination image and keep the illumination of image,at the same time discussed the speed and accuracy of face detection.The main contents of the thesis are as follows:1.Use the traditional histogram equalization to enhance low illumination image is not very well.In order to solve this problem,a novel method is proposed, which combines improved histogram equalization with the contrast enhancement algorithm.The partial information is used to enhance the contrast,and then we adjust the histogram with the improved histogram equalization.Extensive experiments show the efficiency of our method and the computing time remains unchanged.2.A new method is proposed in this paper to solving the problem of color distortion of the color images.Firstly,the histogram is divided of three channels of R,G,and B twice based on the mean of the grey levels and the principle of equal area.Secondly the sub-histograms are equalized respectively.Finally,the processed images of each channel are merged by calculating the proportion of the three channels in original image.The experiments indicate that the proposed method is quite effective in processing color images.3.A new Adaboost algorithm is proposed which have a new threshold option with right update.And this new adaboost algorithm is combined with image enhancement algorithm.A low illumination image is enhanced by the allocation algorithm in this paper at first,and then it is used to detect face by improved Adab0ost algorithm.The experimental result shows that the proposed method can keep the speed of the detection and detect the face in low illumination image effective,at the same time it reduced the false rate of low illumination image in face detection.