Effective Shot Segmentation Method Based on Rough Sets in Video Compressed Domain

Based on rough sets (RS), an effective shot segmentation method is proposed in discrete cosine transform (DCT) compressed domain. First, DCT coefficients and DC coefficients are extracted from video sequences, and an information system is constructed with DC coefficient. Second, information system is reduced by reduction theory of RS, and the representation of the video frames is generated by reduced DC coefficients. Third, by introducing subdividing theory of RS, the frames of video are segmented scientifically. The experimental results show that the method can achieve good performance. Compared to conventional methods, the method enjoys following advantages. Firstly, only a subset of frames needs to be considered during analysis, allowing the reduction of the computational complexity, so the method can avoid the expensive computations in decompression processes. Secondly, the segmentation relativity of video shots becomes more scientific than previous methods. Its robustness and effectiveness are validated by experiments with various kinds of video sequences.