Content based sports video sequences analysis and synthesis

It is very important to analyze and synthesize sports video sequences efficiently for improving competition. Aiming at some limits existing in literature with complex computation and low efficiency in analysis sport video, a novel method has been developed to analyze and synthesize sport videos by intercepting pictures manually or automatically based on visual content. A group of continuous pictures is intercepted automatically from sports video sequences firstly. Some pictures can be added and removed manually based on visual content secondly. Some concerning pictures containing some important semantic information can be gotten in this way. Each picture can be analyzed by some marks added and can be saved in a container. Each group of pictures can be found in the related folder easily and can be demonstrated continuously. Much time is saved and some important semantic information in the sport video can be obtained efficiently by the proposal.