Real Time Face Detection System Based Edge Restoration and Nested K-Means at Frontal View

Bayesian technique is a popular tool for object detection due to its high efficiency. As it compares pixel by pixel, it takes a lot of execution time. This paper addresses a novel framework for head detection with minimum time and high accuracy. To detect head from motion pictures, motion segmentation algorithm is employed. The novelty of this paper carried out with the following steps: frame differencing, preprocessing, detecting edge lines and restoration, finding the head area and cutting the head candidate. Moreover, nested K-means algorithm is adopted to find head location and statistical modeling is employed to determine face or non-face class, while Bayesian Discriminating Features (BDF) method is employed to verify the faces. Finally, the proposed system is carried out with a lot of experiments and a recognizable success is notified.