The Detection Of The Video Advertisement Based On Content In Streaming Media

With the further development of market economy, video advertisementsbecome increasingly important for businesses, and gradually become a part ofroutine work for a company. Advertising companies and enterprises must sentworkers to monitor the implementation of the contract in order to ensure thatvideo advertisements achieve their due effects. At the same time, culturalmanagement department will follow the government’s executive order, requiringtelevision stations to broadcast a number of public service announcement,government notices, notifications, and so on. In order to ensure that these videohave been entirely played within the specified time, cultural managementdepartment must designate persons specially appointed to detect the video.Nowadays advertisements are currently monitored artificially, wasting a lot ofhuman and material resources.In this paper we discuss the characteristics of digital video, the type of shotcut, especially focuse on the key technologies of video detection based on content.In the aspects of the shot boundary detection, we discuss a variety of algorithm,compare the application of these methods, especially use the dual-thresholdmethod to achieve the system, in the aspects of the frame matching, we discuss avariety of frame matching algorithm based on color, compare the application ofeach method, then apply the algorithm based on meaningful regions in the system.We discuss the algorithm, which combine the frame matching with the shotboundary detection, put forward an improved method to improve the detectionperformance of video advertisements. Design and implement the system.Highlight the architecture of the system, and each modular, including: the mainmodular, the modular of the extract of I frame, the modular of shot boundarydetection, the modular of the frame matching. Proved by experiments, the systemcan achieve video advertisements detection commendably.