Temporal video segmentation to scene based on conditional random fileds*

Abstract

In this paper, we propose a novel approach of video segmentation into scenes based on the technique of conditional random fields (CRFs). This approach is built upon the design in which scene segmentation is transformed into a label identification problem by defining three types of shots. To implement our algorithm, three middle-level features including shot difference signal, scene transition graph and audio type are extracted to depict the label properties of each shot, and then CRFs model is employed to identify the labels sequence. The advantage of CRFs model lies in its facility in integrating context information of neighboring shots, which produces accurate results in scene segmentation. The proposed approach is verified by seven types of data covering the most major genres of TV program. Experiments on testing data set yield average 0.88 F-measure, which illustrates that the proposed method can accurately detect most scenes in different genres of programs. © Springer-Verlag 2013.