Key Frames Extraction Algorithm Based on GA

The key frame is a set of discrete frames which contain the most important information in the video. It can greatly reduce the time of video library retrieval by extracting key frames according to an efficient algorithm, and some important information can also be embedded in the key frames to protect the video copyright. In this paper, I propose a new key frame extraction algorithm based on GA (Genetic Algorithm). In this algorithm, a part of frames are thought of as a population, by a series of genetic evolution we can select out the excellent unit as the key frames. The result shows that in a group the frame which greatly differs with others should be selected out as key frame. The algorithm based on GA can select out frames which has more expressive force than traditional algorithm in proper time.