Abstract: Quality Supervision Technology is an important part of IPTV (Internet Protocol Television) and Mobile TV Supervision Technology, it decides the quality which people watch video to some degree. Quality Supervision supervises the quality of transmission and content of quality and quality of users'experience from the level of business quality. This paper does relevant research based on mosaic detection of content supervision. Considering mosaic's the error area and its neighborhood exists not smooth interface, so the detection of mosaic supervises information by the gray image edge of pause coupling neural network. It may extract a very clear rectangle or square shape which is obvious different from other parts by the result of simulation.