AUTOMATIC OBJECTIONABLE VIDEO CLASSIFICATION SYSTEM

摘要

In this paper, we first propose the objectionable video classification system and the novel features that use the skin information of the objectionable videos in order to classify the objectionable videos. First, the proposed system extracts the text information from file header and filename to classify the objectionable video. And then, the proposed features are generated from the video passing text filter and then used as an input to a nonlinear-SVM which determines whether video is objectionable or not. Because the proposed features contain the adequate properties of objectionable video, we expect to improve the performance of the classification system. Experimental results show that the proposed system has an excellent performance in classifying a video as the objectionable or the unobjectionable video.