MSRA-USTC-SJTU AT TRECVID 2007: HIGH-LEVEL FEATURE EXTRACTION AND SEARCH
Tao Mei, Xian‐Sheng Hua, Wei Lai, Linjun Yang, Zheng-Jun Zha, Yuan Liu, Zhiwei Gu, Guo-Jun Qi, Meng Wang, Jinhui Tang, Xun Yuan, Zheng Lu, Jun Liu · 2007
This paper describes the MSRA-USTC-SJTU experiments for TRECVID 2007. We performed the experiments in high-level feature extraction and automatic search tasks. For high-level feature extraction, we investigated the benefit of unlabeled data by semi-supervised learning, and the multi-layer (ML) multi-instance (MI) relation embedded in video by MLMI kernel, as well as the correlations between concepts by correlative multi-label learning. For automatic search, we fuse text, visual example, and concept-based models while using temporal consistency and face information for re-ranking and result refinement. Index Terms — support vector machines, semi-supervised learning, manifold ranking, multi-layer multi-instance kernel, linear neighborhood propagation, temporally consistent Gaussian random field, optimal multi-graph learning, correlative multi-label annotation, video annotation, video search. 1.