Pornographic Video Detection Based On Skin Color And Behavior

The rapid development of Internet and the appearance and quick updates of multimedia technologies have greatly promoted the progress of society as well as international communication. However, they also provide the conditions for pornographic information to emerge and exist at the same time. Pornography has become a major hazard to society, and has seriously affect people’s life, especially for the teen-agers. Compared with pornographic text and images, pornographic videos have made a greater negative impact on teen-agers because of their more realistic and nude scenes. Therefore, to carry out the work for clearing the environment of Internet and screening pornography, have not only a widely theoretical and practical significance for research and production, but also can bring economic and market values. This paper proposes a detection method for pornographic video by a combination of behavior recognition and skin color detection.The behaviors in pornographic videos are mostly relatively reciprocating motion, which is a bright point. The histogram of optical flow can describe the characteristics of the continuous movement. For the above-mentioned reasons, this paper proposes a new feature used to describe the behavior—the optical flow context histogram (OFCH), and uses principal component analysis (PCA) for feature dimension reduction. The PCA-OFCH features are used to train sensitive action classifier. The use of Bayesian skin color prediction model based on histogram techniques can determine if it contains sufficient color information in video, in order to reduce the false alarm rate. Finally we have collected a large number of human pictures, a batch of pornographic videos and a few standard databases for human behavior recognition, for training and testing the sensitive behavior classifier and skin color classifier.Experiments show that the proposed features based on PCA-OFCH with skin color detection can effectively identify pornographic video and normal video, which provides new ideas for pornographic video recognition.