Modification of the Viola-Jones Algorithm for Video Face Identification

Automated video surveillance systems make it possible to compare the image of a human face in several shots and then identify the best shot by analyzing the entire series of them. Today accurate identification of a human face is necessary in a number of industries The most important thing in this case is to search for a face image in a low-resolution video stream and in the presence of various video interferences that make identification difficult. For this reason, this work is devoted to developing an original method for identifying a human face recognized in a series of images. The proposed method has high speed and is able to effectively track the wanted face. As a result, it can be used in electronic identification systems operating online. The identification process within the proposed method is based on the Viola-Jones method which provides comparison of the location of a wanted object in a series of images and subsequent comparison. The results of testing the developed method have demonstrated that the speed of face identification process using it is 37% higher than the speed of the basic unmodified method. Accordingly, the developed modified method can be successfully used to solve the problems of identifying a human face online.