Task-Driven Dynamic Fusion: Reducing Ambiguity in Video Description.

Xishan Zhang, Ke Gao, Yongdong Zhang, Dongming Zhang, Jintao Li, Qi Tian

Browse the full CVPR paper archive.