Broadcast News Video Shot Detection and Segmentation
General trend shows a shift in the way audience consume media. Every information is now communicated in digital form and audiences are interested only in the content that is relevant to them. This is why an efficient video indexing process has become a necessity for news broadcaster to provide the content that its target audience want. In this paper we have proposed an optimized algorithm for shot detection method that aims to reduce the number of false hits that occurs in shot transition detection methods. For this purpose, a video content of a local news channel is taken. Histogram differences and double threshold approaches have been used to detect the abrupt shot transition boundaries. The experimental results showed that the proposed approach is very effective in detecting scene change of large video files with a precision of 81%.
Keywords - Histogram Differences, Abrupt Shot Boundaries