Design And Simulation For A Filtering Pornography System

With the spread and advance of Internet, especially the 3G mobile network, the transmission of information become easy and quick.data can be delivered between two points of internet by kinds of medium.all those make it easy to share data for people. However, Information such as pornography which may harm human's mental health reach to people's eye through the transmission of network unexpectedly and brought to the life of people all sorts of trouble and distress. Filtering pornography is a very important solution to this problem and how to filtering pornography is the main subject of this paper.This paper mainly describes the method of image recognition based on content to reach the goals to filter pornography and design the system for this goal. First, images which are obtained from network should be preprocessed. Some regions of interested will be detected by some detection methods, and those regions can be recognized separately. Normal image can be distinguished by whether information such as skin, human body exist in this image. If human body information exists, those regions will be classified and regional restructure and semantic analysis through the body's geometry, topology and color attribute and finally judge whether the image is pornography. The knowledge of image recognition is introduced, which are preprocessing of image, image segmentation, detection of regions of interesting, analysis of image characteristics, algorithm of pattern recognition and so on. Finally system design is provided and it proved a good performance of this method proposed by this paper. There is a high correct rate while detecting whether a image is a pornography. This system could be making a practical use.