An Effective Shot Segmentation Algorithm Based on Color Feature of DC-Reduced Graph in Compressed-Domain

Shot segmentation is one of the most important steps for video analysis and retrieval. An effective shot segmentation algorithm is proposed by using DC-reduced graph which is on the basis of DC coefficients in I-Frames. This algorithm constructs the sequence of DC-reduced graph by DC coefficients in I-Frame firstly, and then the difference between two adjacent DC-reduced graphs is measured using color moments, spatial distribution entropy of colors and color-histogram. Finally, shot segmentation is achieved through selecting an appropriate threshold. The experiments show that the algorithm has efficiently improved the performance of shot detection, and to a certain extent, the complexity of the detection on gradual shot cuts is reduced.