Abstract
The research aim is apply algorithm for finding the number of regions to determine the largest number of faces of candidates in the video or detection, which appear in the video depending on the color techniques in the video and to fragment the skin and reject the largest number of regions that are non skin, and reduce the false faces.
The algorithm depends on the usage of technique to detect human face and select it in the video after segment it to set of color images, the technique includes two steps for image processing The first is for building a model for skin color, and an outline of the distribution of color and isolate areas of skin and non skin within the image, then identify areas of the skin. The second step includes the template matching.
The results proved high accuracy by nearly 85% in the discrimination of the faces in the video and enclosed in geometric and stored in a new video, the algorithms is programmed in MATLAB 7.10.0 (2010) programming language which has many efficient functions in image processing.
