Video shot boundary detection by dominant sets approach

In this study, a video shot boundary detection algorithm based on the dominant sets concept is proposed. Dominant sets method is a graph theoretic clustering algorithm. Proposed method is based on a weighted undirected graph. Candidate shot boundaries are determined and graphs are constructed by taking 2 frames from the right of the candidate position and 4 frames from the left of the candidate position. Edge weights among the vertices are evaluated by using pairwise similarities of frames. By using the complete information of the graph, a set of the vertices mostly similar to each other and dissimilar to the others is detected. True cut positions are determined if the dominant set includes the 4 frames before the candidate position. The simulation results indicate that the proposed algorithm can be used for abrupt shot boundary detection.