Classifying documentary, music, news and animated genres with temporal, color and contour information

In this paper we address the issue of automatic video genre classification and propose three categories of content descriptors. At temporal level, video content is described in terms of visual rhythm, action content and amount of gradual transitions. Further, colors are globally described using statistics of color distribution, elementary hues, color properties and relationship. Finally, structural information is extracted at image level and histograms are built to describe overall contour features and their relations. The proposed descriptors were used to classify four of the most common video genres, thus: animated movies, documentaries, music clips and newscast. Experimental tests conducted on more than 67 hours of video footage prove the high efficiency of these features. We achieve an average correct detection ratio up to 95%, while the precision and recall ratios are up to 98% and 100%, respectively.