Multiple face detection in color images

Abstract

Face is the primary index for imparting the identity. Automated face detection is one of the interesting field of research. Face detection of digital image has acquired much importance and interest in last two decades, which has applications in different fields. Computerizing the process needs many image processing methods. In this paper, a new face detection approach using color base segmentation and morphological operations is presented. The algorithm uses color plane extraction, background subtraction, thresholding, morphological operations (such as erosion and dilation), filtering (to avoid false detection). Then particle analysis is done to detect only the face area in the image and not the other parts of the body. The color planes are extracted using vision module the RGB color space is converted into suitable color space such as HSV and YCbCr. The algorithm can be used to detect both single as well as multiple persons in a image. Experimental results of the algorithm show that, it is good enough to detect the human faces with an accuracy of 93% i.e., the efficiency of the detection is up to 93%.