Abstract HTML Views: 471 PDF Downloads: 196 Total Views/Downloads: 667
Abstract HTML Views: 299 PDF Downloads: 148 Total Views/Downloads: 447
Static video abstract is able to show the dynamic event of semantics. According to the current research of
events, the method of generating speaking video summary by detecting face in moving region is proposed. Extracting
face-based frame and its moving region, combined with the trained face classifier, the speaking event is detected. Integrated
video temporal and spatial characteristics, this method expresses the visual impact of video contents effectively.
Experimental results indicate the generated summary with good effects.