Improving the video shot boundary detection using the HSV color space and image subsampling

High speed video analysis often requires efficient shot boundary detection to break a video into meaningful continuous segments. In this paper we propose new techniques to improve the efficiency in terms of both processing speed and detection accuracy upon the state-of-the-art methods. Specifically, we use the HSV color space in stead of the traditional RGB color space. This turns out to produce more robust detection results. In addition, each image frame is subsampled based on which a frame distance is defined. It turns out that such processing significantly lowers the computational complexity while yielding the same level of detection accuracy. Simulation results confirm the advantages of the proposed techniques over the state-of-the-art methods.