A New Algorithm of Counting Human Based on Segmentation of Human Faces in Color Image

The face detection problem has attracted wide attention. Some achievements have been obtained in the past decade. The paper presented a novel algorithm of face detection based on segmentation of human faces in color image. The number of human can be counted and can be used to control the lights in a classroom to be all turned on or not based on proposed method in the paper. First, this paper constructs a mixed model in both RGB (Red, Green, Blue) and HSV (Hue, Saturation, Value) color space, which can be used to acquire possible skin areas in original image. Second, non-face objects such as human hands and arms in the image are eliminated with the character of face contour. Thus, the candidate faces are extracted. Third, confirmed faces can be extracted by using the color and shape characteristics of hair. Experiment result shows this algorithm is efficient and available.