AI Paper Atlas

Research map / Video / Video segmentation and tracking

Motion-Attentive Transition for Zero-Shot Video Object Segmentation

Tianfei Zhou, Shunzhou Wang, Yi Zhou, Yazhou Yao, Jianwu Li, Ling Shao

AAAI 2020 · Video › Video segmentation and tracking · approach: Video semantic and panoptic segmentation · 188 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 motion-attentive transitions between motion and appearance features at each encoder stage (ablation)MATNet · DAVIS-16Mean J: 79.5 -> 82.4Improved
Add motion-attentive transitions between motion and appearance features at each encoder stage (ablation)MATNet · DAVIS-16Mean F: 77.3 -> 80.7Improved
Use SSA rather than a simple skip layer (ablation)MATNet · DAVIS-16Mean J: 80.7 -> 82.4Improved
Use SSA rather than a simple skip layer (ablation)MATNet · DAVIS-16Mean F: 79.7 -> 80.7Improved
Add hard example mining for boundary training (ablation)MATNet · DAVIS-16Mean F: 78.4 -> 80.7Improved
Add hard example mining for boundary training (ablation)MATNet · DAVIS-16Mean J: 81.4 -> 82.4Improved
Replace ResNet-101 with ResNet-50 (ablation)MATNet · DAVIS-16Mean J: 82.4 -> 81.1Hurt
Increase MAT depth from L=0 to L=1 (ablation)MATNet · DAVIS-16Mean J: 79.5 -> 80.6Improved
Increase MAT depth from L=1 to L=3 (ablation)MATNet · DAVIS-16Mean J: 80.6 -> 81.6Improved
Increase MAT depth from L=3 to L=5 (ablation)MATNet · DAVIS-16Mean J: 81.6 -> 82.4Improved
Increase MAT depth from L=5 to L=7 (ablation)MATNet · DAVIS-16Mean J: 82.4 -> 82.2Hurt
Use the interleaved MATNet encoder rather than independently learned motion and appearance streamsZero-shot video object segmentation · DAVIS-16 valMean J: 80.7 -> 82.4Improved

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