AI Paper Atlas

Research map / Video / Video segmentation and tracking

Crossover Learning for Fast Online Video Instance Segmentation

Shusheng Yang, Yuxin Fang, Xinggang Wang, Yu Li, Chen Fang, Ying Shan, Bin Feng, Wenyu Liu

ICCV 2021 · Video › Video segmentation and tracking · approach: Mask propagation for unsupervised video segmentation · 105 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 crossover learning (COL) (ablation)CondInst-VIS · YouTube-VIS-2019 valAP: 32.1 -> 33.1Improved
add global balanced embedding (GBE) (ablation)CondInst-VIS · YouTube-VIS-2019 valAP: 32.1 -> 33.5Improved
add both COL and GBE (ablation)CondInst-VIS · YouTube-VIS-2019 valAP: 32.1 -> 34.8Improved
add crossover learning with large sample interval T=20 (ablation)CrossVIS · YouTube-VIS-2019 valAP: 33.4 -> 34.8Improved
add crossover learning with small sample interval T=1 (ablation)CrossVIS · YouTube-VIS-2019 valAP: 33.1 -> 33.6Improved
use global instead of pair-wise embedding (CE loss) (ablation)instance association · YouTube-VIS-2019 valAP: 33.1 -> 33.4Improved
use focal loss instead of CE for pair-wise embedding (ablation)instance association · YouTube-VIS-2019 valAP: 33.1 -> 33.3No clear effect
use global embedding with focal loss instead of pair-wise CE (ablation)instance association · YouTube-VIS-2019 valAP: 33.1 -> 34.8Improved
use global instead of pair-wise embedding for stability (CE) (ablation)instance association · YouTube-VIS-2019 valAP std: 0.78 -> 0.27Improved
add multi-scale trainingCrossVIS · YouTube-VIS-2019 valAP: 34.8 -> 36.3Improved
add temporal feature calibrationCrossVIS · OVIS valAP: 14.9 -> 18.1Improved
use ResNet-101 instead of ResNet-50CrossVIS · YouTube-VIS-2019 valAP: 34.8 -> 36.6Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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