A system for the semantic multimodal analysis of news audio-visual content
EURASIP Journal on Advances in Signal Processing
Pattern Recognition Letters
Hi-index | 0.00 |
Video comprises multiple types of textual, audio and visual information, and each of them contains abundant semantic information. Therefore multimodal features query and fusion are necessary in video retrieval. In this paper, we propose a new video retrieval model, which adopts multi-model including text, image, semantic concept and camera motion to query video. Then relation algebra expression is advanced to fuse multimodal information instead of traditional linear fusion method. In semantic concept detection model, Bayesian network based ontology is proposed to extract concepts. The experiments on TRECVID 2005 corpus have demonstrated a superior performance compared with exiting key approaches of video retrieval by multimodal information fusion.