Classification based on SVM of cultural relic videos´ key frame

For the question of how to classification of cultural relic videos, this paper put forward the analysis method based on semantic and it´s two-steps in classification of video: Firstly, based on algorithm that has been used, separating the cultural relic shots and extracting the key frames; Then extracting the features of the key frames. Next training classification of the key frames´ features using SVM. The system composed by the two steps´ algorithms implementing classification of cultural relic video. Experiments show this method can better classify based on the characteristics of cultural relic videos then other methods.