CLIP2TV: Align, Match and Distill for Video-Text Retrieval

10 Nov 2021  ·  Zijian Gao, Jingyu Liu, Weiqi Sun, Sheng Chen, Dedan Chang, Lili Zhao ·

Modern video-text retrieval frameworks basically consist of three parts: video encoder, text encoder and the similarity head. With the success on both visual and textual representation learning, transformer based encoders and fusion methods have also been adopted in the field of video-text retrieval. In this report, we present CLIP2TV, aiming at exploring where the critical elements lie in transformer based methods. To achieve this, We first revisit some recent works on multi-modal learning, then introduce some techniques into video-text retrieval, finally evaluate them through extensive experiments in different configurations. Notably, CLIP2TV achieves 52.9@R1 on MSR-VTT dataset, outperforming the previous SOTA result by 4.1%.

PDF Abstract

Results from the Paper


Ranked #5 on Video Retrieval on MSR-VTT (using extra training data)

     Get a GitHub badge
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Video Retrieval MSR-VTT CLIP2TV text-to-video R@1 33.1 # 5
text-to-video R@5 58.9 # 6
text-to-video R@10 68.9 # 5
text-to-video Mean Rank 44.7 # 3
text-to-video Median Rank 3 # 1
Video Retrieval MSR-VTT-1kA CLIP2TV text-to-video Mean Rank 12.8 # 5
text-to-video R@1 52.9 # 6
text-to-video R@5 78.5 # 4
text-to-video R@10 86.5 # 4
text-to-video Median Rank 1 # 1
video-to-text R@1 54.1 # 3
video-to-text R@5 77.4 # 3
video-to-text R@10 85.7 # 3
video-to-text Median Rank 1 # 1
video-to-text Mean Rank 9.0 # 5

Methods


No methods listed for this paper. Add relevant methods here