Lighting change problem in shot detection

A lot of work has been done to address shot detection, e.g. pixel-by-pixel comparisons and histogram comparisons. But gray or color histograms ignore the problem caused by lighting condition variations inherent in the video production process especially in TV news reports. So they often can be confused when such incidents occur. Color Ratio Histogram is proposed to address the issue. However it is rather computationally expensive. Three methods are proposed in this paper. The first one takes the advantage of the lasting time of the illumination variation and the second one takes into consideration the regulation of the change of gray level of blocks in the consecutive images due to lighting condition variation. The last uses another illumination invariant metric to deal with the problem. The effectiveness of the methods is validated by experiments on some real-world video sequences