Research On Adaptive Detection Of Progressive Film Mode

Television(TV) is moving in the digital high-definition direction after a century of innovation. In the era of analog TV, due to the bandwidth constraint, TV displayed video employing the interlaced scanning. In order to play the progressive film in the interlaced TV, it was necessary to transform the progressive film into the interlaced one using the pulldown technology. With the development of technology, TV entered into the digital age and could play progressive sources directly. To improve the picture quality of the displayed video, the de-interlacing technique was introduced to transform the interlaced sources into the progressive sources. However, the de-interlacing technique does not identify and remove the duplicate fields brought about by the pulldown technology, so there are lots of redundant frames in the progressive sources which lead to the video jitter. To solve this problem, the redundant frames in the progressive sources should be detected and removed to restore the original film, where the progressive film mode detection is the key technology.The Characteristics of the different types of progressive film sources are discussed deeply, and the limitations of the existing detection algorithms are analyzed in this paper. By summarizing the characteristics of the known film source modes, the parameters of four preset sliding windows are determined, and a sliding-window-based adaptive progressive film mode detection algorithm is proposed. The proposed algorithm analyzes the difference between the key frame and the duplicate frame, and then utilizes the relationship between the inter-frame difference and the film source characteristic to adaptively determine the optimal window size reflecting the characteristics of the film to test. The different types of progressive film sources can be detected according to the corresponding relationship between the key frame proportion in the sliding window and the characteristic of the original film source. The proposed adaptive algorithm breaks the limitations of the existing detection algorithms and can detect accurately both the de-interlaced progressive films and the self-processing films of the player.To further extend the application of the progressive film detection algorithm, the progressive film sources containing scene changes are analyzed in this paper. Since the scene change may damage the rule of the progressive film sources, the accuracy of the scene change detection algorithm will directly affect the performance of the progressive film mode detection. By analyzing the advantages and disadvantages of the existing scene change detection algorithms, a scene change detection algorithm is proposed combining the inter-frame difference and the block-based chi-square histogram matching method. According to the frame rate changes before and after scene changes, a detection algorithm for the progressive film containing scene changes is proposed. If the frame rate doesn't change after the scene change, the proposed algorithm maintains the current locked mode; on the contrary, the proposed algorithm quits the current locked mode and detects the film source after scene changes again. The proposed algorithm pre-determines the original frame rate by checking whether the key frame proportion in the sliding window changes or not before and after scene changes, and identifies the film sources after scene changes quickly.