AI Paper Atlas

Research map / Video / Video segmentation and tracking

TCOVIS: Temporally Consistent Online Video Instance Segmentation

Junlong Li, Bingyao Yu, Yongming Rao, Jie Zhou, Jiwen Lu

ICCV 2023 · Video › Video segmentation and tracking · approach: Matching-based video object segmentation · 32 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
Replace local matching and propagating with global instance assignment (ablation)TCOVIS ResNet-50 · YouTube-VIS 2021AP: 47.4 -> 48.6Improved
Add spatio-temporal enhancement module on top of global assignment (ablation)TCOVIS ResNet-50 · YouTube-VIS 2021AP: 48.6 -> 49.5Improved
Add both GIA and STE over LocPro baseline (ablation)TCOVIS ResNet-50 · YouTube-VIS 2019AP: 50.6 -> 52.3Improved
Use cross-attention with shared positional embedding instead of concatenation for STE (ablation)STE · YouTube-VIS 2021AP: 47.9 -> 49.5Improved
Add self-attention layer before cross-attention in STE (ablation)STE · YouTube-VIS 2021AP: 49.5 -> 48.2Hurt
Apply enhancement at every decoder layer (decoder modulation) (ablation)STE · YouTube-VIS 2021AP: 49.5 -> 48.5Hurt
Share positional embedding between spatial feature and query of same instance (ablation)STE · YouTube-VIS 2021AP: 48.8 -> 49.5Improved
Decode spatial features of all instances rather than only the instance itself (ablation)STE · YouTube-VIS 2021AP: 49.2 -> 49.5Improved
TCOVIS versus online GenVIS with Swin-LTCOVIS · YouTube-VIS 2022AP: 45.1 -> 51.0Improved
TCOVIS versus online GenVIS with ResNet-50TCOVIS · OVISAP: 35.8 -> 35.3Hurt
TCOVIS versus IDOL with Swin-LTCOVIS · YouTube-VIS 2019AP: 64.3 -> 64.1No clear effect

Later papers that compared with this method

1 later accepted paper reports TCOVIS as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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