Papers › Lightweight Attentional Feature Fusion: A New Baseline for Text-to-Video Retrieval

Lightweight Attentional Feature Fusion: A New Baseline for Text-to-Video Retrieval

3 Dec 2021arXiv:2112.01832archive 2025-07-28

Fan Hu, Aozhu Chen, Ziyue Wang, Fangming Zhou, Jianfeng Dong, Xirong Li

In this paper we revisit feature fusion, an old-fashioned topic, in the new context of text-to-video retrieval. Different from previous research that considers feature fusion only at one end, let it be video or text, we aim for feature fusion for both ends within a unified framework. We hypothesize that optimizing the convex combination of the features is preferred to modeling their correlations by computationally heavy multi-head self attention. We propose Lightweight Attentional Feature Fusion (LAFF). LAFF performs feature fusion at both early and late stages and at both video and text ends, making it a powerful method for exploiting diverse (off-the-shelf) features. The interpretability of LAFF can be used for feature selection. Extensive experiments on five public benchmark sets (MSR-VTT, MSVD, TGIF, VATEX and TRECVID AVS 2016-2020) justify LAFF as a new baseline for text-to-video retrieval.

PaperPDFCode

In Syntology View this paper on Syntology: its repositories, every harvested function with whether it ran, its licence and the call to fetch it.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

ruc-aimc-lab/laff officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Ad-hoc video searchRetrievalText to Video RetrievalVideo Retrievalfeature selection

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Ad-hoc video search TRECVID-AVS16 (IACC.3) LAFF infAP 0.222 #1 of 4 Archive leaderboard report
Ad-hoc video search TRECVID-AVS17 (IACC.3) LAFF infAP 0.290 #1 of 4 Archive leaderboard report
Ad-hoc video search TRECVID-AVS18 (IACC.3) LAFF infAP 0.147 #1 of 4 Archive leaderboard report
Ad-hoc video search TRECVID-AVS19 (V3C1) LAFF infAP 0.192 #1 of 2 Archive leaderboard report
Ad-hoc video search TRECVID-AVS20 (V3C1) LAFF infAP 0.265 #1 of 1 Archive leaderboard report
Video Retrieval MSR-VTT LAFF text-to-video R@1 29.1 #27 of 40 Archive leaderboard report
Video Retrieval MSR-VTT LAFF text-to-video R@10 65.8 #27 of 40 Archive leaderboard report
Video Retrieval MSR-VTT LAFF text-to-video R@5 54.9 #27 of 40 Archive leaderboard report
Video Retrieval MSR-VTT-1kA LAFF text-to-video R@1 45.8 #35 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA LAFF text-to-video R@10 82 #35 of 63 Archive leaderboard report
Video Retrieval MSR-VTT-1kA LAFF text-to-video R@5 71.5 #35 of 63 Archive leaderboard report
Video Retrieval MSVD LAFF text-to-video R@1 45.4 #20 of 24 Archive leaderboard report
Video Retrieval MSVD LAFF text-to-video R@10 84.6 #20 of 24 Archive leaderboard report
Video Retrieval MSVD LAFF text-to-video R@5 76.0 #20 of 24 Archive leaderboard report
Video Retrieval TGIF LAFF text-to-video R@1 24.5 #2 of 2 Archive leaderboard report
Video Retrieval TGIF LAFF text-to-video R@10 54.5 #2 of 2 Archive leaderboard report
Video Retrieval TGIF LAFF text-to-video R@5 45.0 #2 of 2 Archive leaderboard report
Video Retrieval VATEX LAFF text-to-video R@1 59.1 #11 of 13 Archive leaderboard report
Video Retrieval VATEX LAFF text-to-video R@10 91.7 #11 of 13 Archive leaderboard report
Video Retrieval VATEX LAFF text-to-video R@50 96.3 #11 of 13 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections