基于镜头竞争力的多模态视频场景分割算法

Multi-Modality Video Scene Segmentation Algorithm Based on Shot Force Competition

为了能快速、有效地进行视频场景分割,论文提出一种基于镜头竞争力的多模态视频场景分割算法,充分考虑视频中多模态之间的时序关联共生特性,通过对视频物理特征的提取、融合计算出镜头间相似度,结合镜头竞争力的判定思想分割出视频场景。实验结果表明,该算法能较为高效地进行视频场景分割,查全率和查准率可达82.1%和86.7%。

In order to quickly and effectively video scene segmentation,a multi-modality video scene segmentation algo rithm based on shot force competition is proposed realize.Take full account of temporal associated co-occurrence of multimodal media data in video,the video shot similarity is calculated by extracting and merging the video physical features,and shot competitive judgment method is combined to conduct video scene segmentation.The experiments show that the video scene can be effectively separated by the proposed method,and recall,precision reach 82.1 ％,86.7％ respectively.