Content-based drama editing based on inter-media synchronization

Abstract

Automatic image recognition is usually difficult. However, drama is composed of not only image but also sound and script. Therefore, higher level recognition of drama which cannot be achieved by the image analysis will be realized by integrating image analysis, sound analysis and document analysis. In this paper, focusing on the semantic descriptions in the script, content-based drama editing based on inter-media synchronization is proposed. In the proposed method, inter-media synchronization between image, sound and script is carried out using DP matching. Then, the script is edited using an ordinary text editor. Finally, the editing of drama is carried out by selecting and aligning the image and sound based on the edited script. Therefore, the proposed method can realize effective and semantic-based video editing by handling smaller and more abstracted data. The experimental results reveal the effectiveness of the proposed method.