Study And Analysis On Video Matching Based On Ts Model

As more and more multimedia videos storage in the network, it makes the requirements of video matching more and more strict. So how to match videos quickly and accurately becomes a hot issue. In recent years, with the ripeness of video processing technologies, more and more video matching technology began to be fully applied in the field of medicine, television broadcasting system, as well as traffic monitoring system. It has greatly improved a lot of duplication of multimedia video database, and provided technical support to network database. However, in practical applications, the blindness and imbalance of the video matching technology become more and more serious.In order to overcome the drawbacks of more mistakes in the existing video matching methods, and the maladjustment in multi-shots video, this paper introduces the model of Telescopic-shot, and use the method of video shot detection, and separate the large video into short videos, to get the accurate matching sequences. Finally reduces the time complexity of large video. By comparing with the existing methods, the experimental results show, the time complexity of this paper is significantly reduced, and the rate of mistake is lower.In order to overcome the shortcomings of video matching methods, make up for the inaccuracy of video matching based on motion features and improve the large computing complexity of video matching based on content, and so on, this paper will put forward a TS method based on dominant color and texture feature. It will improve the efficiency of video matching, from the perspective of reducing time complexity and increasing the registration rate. By many experiments, this method solves the above problems effectively. Registration rate increases by about20%than traditional methods, and the time complexity of algorithm reduces by about45%.