AI Paper Atlas

Research map / Video / Video-language understanding

Enhancing Video-Language Representations With Structural Spatio-Temporal Alignment

Hao Fei, Shengqiong Wu, Meishan Zhang, Min Zhang, Tat‐Seng Chua, Shuicheng Yan

TPAMI 2024 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 72 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Finsta structural scene-graph alignmentInternVideo · K400Top-1 accuracy: 91.1 -> 93.7Improved
add FinstaHDVILA · MSR-VTT captioningMETEOR: 32.4 -> 36.9Improved
add FinstaHDVILA · LSMDC retrievalR@1: 17.4 -> 25.3Improved
add FinstaVideo-LLaVA · MSVD-QAaccuracy: 65.7 -> 72.5Improved
add FinstaAll-in-one · MSR-VTT captioningMETEOR: 32.7 -> 34.2Improved
add FinstaLFVILA (L-Vid) · ActivityNet video-paragraph retrievalR@1: 35.3 -> 69.8Improved
add FinstaLFVILA (L-Vid) · How2QAaccuracy: 76.1 -> 84.8Improved
add Finsta in zero-shot instead of fine-tuned settingVideo-LLaVA · MSVD-QA zero-shotaccuracy: 25.1 -> 45.3Improved
add Finsta in zero-shotInternVideo · K400 zero-shotTop-1 accuracy: 53.3 -> 68.9Improved
add temporal coreference edges in the dynamic scene graph (ablation)Finsta · 6-task averageaverage score: 59.2 -> 62.9Improved
add adverbial modifier nodes in the textual scene graph (ablation)Finsta · 6-task averageaverage score: 60.0 -> 62.9Improved
add cross-modal coreference edges in the holistic scene graph (ablation)Finsta · 6-task averageaverage score: 60.5 -> 62.9Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy