Research On Graphic Bad Content Detection Technology

With the development of triple-play service for many years, it makes life moreConvenient. but the network will become more open and complex, And at the same time,the number of information on the network and the speed of it spread will increase quickly,and the information contained with pornographic, violent and other illegal informationwill have more channels to spread, and spreading wider and wider, the harm will be moreserious. In order to ensure the safety of TV content, We need a system to ensure theauthenticity, efficacy and safety of the content. This system must provide a multi-level ofcontent-monitoringGraphic Bad Content Detection Technology is a very important technology to theContent Supervision System base on the triple play network. Bad content can be prevent totransfer in network with help of the Bad Content Detection Technology. This technologycan be also used in the image retrieval system and another system which need contentunderstanding.At first, after a lot of investigation and research, appropriate definition of BadContent can be found, Bad Content refer to two aspect one is law and the other is morality.The content with sex and bloody violence are not allowed to exist in the network.Based on this, the Process of Graphic Bad Content Detection can be designed, inorder to adjust this system, some algorithm is improved. This technology is based on thevisual features of the image, in the way the feature of sex and bloody violence can bedetected. With the help of The Kohonen neural network and Back Propagation neuralnetwork, the content with sex and bloody violence can be recognized.