Segmenting video shots using 2D entropy model

A shot presents a contiguous action recorded by an uninterrupted camera operation and frames within a shot keep spatio-temporal coherence. Segmenting a serial video stream file into meaning shots is the first pass for the task of video analysis, content-based video understanding. In this paper, a hierarchical framework based on an improved two-dimensional entropy model is proposed to complete the purpose of the partition of video shots. Our approach first selects the transition boundary candidates in the state of desampling. Then these detected transition candidates are further classified into cut or gradual type, and at the same time the falsely detected shot breaks can be distinguished and merged into the normal shot based on the human perception principle of visual content. Moreover, the boundary of gradual transition can be precisely located using characteristic features of gradual transition. A large number of video sequences are used to test our system performance and promising results have been obtained.