AI Paper Atlas

Research map / Video / Video segmentation and tracking

Deformable Video Transformer

Jue Wang, Lorenzo Torresani

CVPR 2022 · Video › Video segmentation and tracking · approach: Video restoration with temporal priors · 43 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
Replace fixed MHPA with D-ST+MS-AMViT-B · Kinetics-400Top-1 accuracy: 78.4% -> 79.0%Improved
Replace trajectory attention with D-ST+MS-AViT-B · Kinetics-400, 16×224² clipsTop-1 accuracy: 79.7% -> 80.5%Improved
Replace MHPA with D-ST+MS-AMViT-B · Something-Something-V2Top-1 accuracy: 64.7% -> 67.5%Improved
Replace trajectory attention with D-ST+MS-AViT-B · EPIC-KITCHENS-100 Action, 16×224² clipsaccuracy: 43.1% -> 43.8%Improved
Replace divided space-time attention with D-ST+MS-AViT-B · Diving-48, same input resolutionTop-1 accuracy: 78.0% -> 86.6%Improved
Use motion displacements rather than averaged pooled RGB, RGB residuals, or RGB differences (ablation)DVT D-ST-A · Kinetics-400 and Something-Something-V2accuracy: K400: 72.3% -> 77.9%; 73.2% -> 77.9%; 73.6% -> 77.9%. SSv2: 62.9% -> 65.6%; 63.3% -> 65.6%; 63.5% -> 65.6%.Improved
Use optical flow instead of motion displacements (ablation)DVT D-ST-A · Kinetics-400 and Something-Something-V2accuracy: K400: 77.9% -> 78.2%; SSv2: 65.6% -> 66.0%.Improved
Add multi-scale deformable attention to space-time deformable attention (ablation)DVT MViT-B · Kinetics-400accuracy: 77.9% -> 78.1% (linear); 77.9% -> 79.0% (MLP-Mixer)Improved
Use MLP-Mixer instead of linear projection for attention fusion (ablation)DVT D-ST+MS-A, MViT-B · Kinetics-400accuracy: 78.1% -> 79.0%Improved
Use MLP-Mixer instead of linear projection for attention fusion (ablation)DVT D-ST+MS-A, MViT-B · Something-Something-V2accuracy: 66.8% -> 67.5%Improved
Increase attended patches from N=4 to N=8 (ablation)DVT D-ST+MS-A · Kinetics-400 and Something-Something-V2accuracyImproved
Increase the number of sub-clips to B=8 (ablation)DVT D-ST-A · Something-Something-V2accuracyHurt

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