Video Object Segmentation Based On Genetic Algorithm

With the fast development of the multimedia technique, the application of video has already been gradually becoming research emphases in the field of the current multimedia technology. The technique of the video object segmentation is the base of the various technique of video processing, and the accuracy of the segmentation will directly affect the valid of fellow-up operations. Therefore, the technique of the video object segmentation have the very important meaning. But the segmentation is one of the most difficult problems to deal. Although domestic and international scholars put forward many methods, up to now there is no one which is wildly used and reliable.This paper further studied the technique of the video object segmentation on the foundation of analyzing the structure of video sequence. First, basic theories on technique of video object segmentation are reviewed and analyzed systematically, including the technique of the image segmentation and mathematical morphology segmentation, and a new method of threshold segmentation is proposed. Then the genetic algorithm is talked about; finally status quo, classifications of the technique and basic methods are discussed subsequently.This study launches the research aiming at video object segmentation based on summarizing and analyzing the domestic and international related research work. The main contents and research production include following aspects: adopting the method of background to distill the moving district of the foreground object; using the new threshold method in order to enhance the definition while segmenting the video object; using improved genetic algorithm to reach the optimization. The result of the experiment show more clear definition of the video object when using the method mentioned in this paper compared with the traditional method. The improved genetic algorithm mentioned in this paper can shorten the time that is cost by searching the best threshold value, improve the efficiency of running program and consequently mend the speed of processing the video segmentation.