Clip-based similarity measure for hierarchical video retrieval

This paper proposes a new approach and algorithm for the similarity measure of video clips. The similarity is mainly based on two bipartite graph matching algorithms: maximum matching (MM) and optimal matching (OM). MM is able to rapidly filter irrelevant video clips, while OM is capable of ranking the similarity of clips according to the visual and granularity factors. Based on MM and OM, a hierarchical video retrieval framework is constructed for the approximate matching of video clips. To allow the matching between a query and a long video, an online clip segmentation algorithm is also proposed to rapidly locate candidate clips for similarity measure. The validity of the retrieval framework is theoretically proved and empirically verified on a video database of 21 hours. Categories and Subject Descriptors: H.3.3 [Information Search and Retrieval]: Retrieval models. General Terms: Algorithms, Experimentation, Theory.

• Type: Conference
• Year: 2004
• Where: MIR
• Authors: Yuxin Peng, Chong-Wah Ngo
• Keywords: MIR 2004, Multimedia, Retrieval Framework, Similarity Measure, Video Clips