An improved shot segmentation algorithm based on color histograms for decompressed videos

This paper proposes an improved shot segmentation algorithm based on color histograms for decompressed videos. Firstly, we quantify the HSV color space non-uniformly and divide the images into 9 different sub-blocks, to which different weight values are set. Then, we calculate the color histogram of each sub-block in the HSV color space and compare the histogram differences of adjacent frames with the adaptive threshold values in order to identify the switching frames. The experimental results demonstrate that through the proposed algorithm we can obtain higher recall and precision rates.