An Approach to Video Key-frame Extraction Based on Rough Set

Since video key frames are pictures that can represent main content of the shot, the key-frame extraction has been the foundation of video analysis and content-based retrieval. At present, most existing key-frame extraction techniques are based on original video streams. Due to the time-consuming procedure of decoding, those methods are obviously inefficient. In this paper, we extract motion information from compressed MPEG streams and apply rough set theory to obtain the decision rules implicated in motion information and key frames. Experimental results show that this approach is able to represent the visual content with less computational time.