Affective Content-Based Film Clips Retrieval Algorithm Using Improved Fuzzy Comprehensive Evaluation

With the development of computer´s communication and multimedia technologies, the amount of video data is enlarged at an explosive speed. How to realize the personal video service becomes a hot issue in video retrieval area. In this paper, a new retrieval approach is proposed base on the scene affective content evaluation. First, affective content evaluation values are computed by improved fuzzy comprehensive evaluation model corresponding to different emotion types, and they are combined to become a vector which is used to describe the content of every scene. Then a film clip retrieval system is established by using triangle similarity measure. Finally, the experiment results are given. The experimental results show that the proposed method can perform best with a high retrieval accuracy rate (exceeding 88%).