Epitome key information extraction using color values on block

Abstract

In this paper, we propose block diagonal movement technique to classify shots on video frame sequence. Both RGB color pixel values and Hue value are employed in the proposed technique. Since Key frames are essential to analysis on large amount of video sequence databases this paper uses surveillance video files to extract some meaningful information key frames from long video sequences. This purpose is to reduce weak transaction frames in a large video stream by using block base approaches with diagonal movement. The experimental results show that RGB pixel value base approach is more suitable than HSV hue value base approach. In addition, it is learnt that RGB pixel value base method can extract key frames than HSV hue base with less processing times. RGB pixel values base diagonal block method can process accurately, clearly and stability to extract key information frames.