Efficient Scene Change Detection and Camera Motion Annotation for Video Classification

With the fast growth of video sources, efficient video classification and management is becoming more and more important. Video partitioning and video feature extraction are two of the key issues in video classification. In this paper, we introduce our innovative approaches to scene change detection and camera motion extraction.The video partitioning process involves the detection of boundaries between uninterrupted segments (video shots) of scenes. Shot boundaries can be classified into two categories, gradual transition and instantaneous change (called a camera break). Detection of a gradual transition is considered to be a difficult problem. Few methods have been reported for gradual transition detection. We discuss an efficient method which is calledStep-variable.In this paper, a novel method is derived to classify the dominant camera motions in video shots. The method is to analyze the optical flow in a decomposed manner. Images are divided into subregions according to our camera model. The projectedxandycomponents of optical flow are analyzed separately in the different subregions of the images. Different camera motions are recognized by comparing the computed result with the prior known patterns. Our method is efficient and effective because only some mean values and standard deviations are used. The analysis and a detailed description of our method are presented in this paper. Experimental results are given to show the effectiveness of the proposed method.