AI Paper Atlas

Research map / Video / Video segmentation and tracking

Learning Pixel-Level Distinctions for Video Highlight Detection

Fanyue Wei, Biao Wang, Tiezheng Ge, Yuning Jiang, Wen Li, Lixin Duan

CVPR 2022 · Video › Video segmentation and tracking · approach: Weakly supervised video anomaly detection · 30 citations (OpenAlex)

Check 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 temporal context from preceding frames (ablation)PLD-VHD with C3D · YouTube HighlightsmAP: 0.651 -> 0.712Improved
Add saliency masks to pseudo-label generation (ablation)PLD-VHD with C3D · YouTube HighlightsmAP: 0.664 -> 0.712Improved
Add temporal context from preceding frames (ablation)PLD-VHD with TASED-Net · YouTube HighlightsmAP: 0.667 -> 0.730Improved
Add saliency masks to pseudo-label generation (ablation)PLD-VHD with TASED-Net · YouTube HighlightsmAP: 0.702 -> 0.730Improved
Add temporal context from preceding frames (ablation)PLD-VHD with TASED-Net · TvSumtop-5 mAP: 0.729 -> 0.771Improved
Add saliency masks to pseudo-label generation (ablation)PLD-VHD with TASED-Net · TvSumtop-5 mAP: 0.741 -> 0.771Improved
Add temporal context from preceding frames (ablation)PLD-VHD with TASED-Net · CoSumtop-5 mAP: 0.888 -> 0.946Improved
Add saliency masks to pseudo-label generation (ablation)PLD-VHD with TASED-Net · CoSumtop-5 mAP: 0.915 -> 0.946Improved
Use TASED-Net rather than C3D as the backbone (ablation)PLD-VHD · YouTube HighlightsmAP: 0.712 -> 0.730Improved

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