AI Paper Atlas

Research map / Label-efficient and robust learning / Test-time adaptation

Mixture of Prototypes for Test-time Adaptive Segmentation

Guangrui Li, Zhengyu Zhu, Yongxin Ge

CVPR 2026 · Label-efficient and robust learning › Test-time adaptation · approach: Pseudo-label test-time adaptation

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 expert filtering (top-K) to L_min-trained gating (ablation)MoE prototypes · Cityscapes-to-ACDCAvg mIoU: 57.9 -> 61.7Improved
optimize gating with entropy minimization L_min alone (ablation)MoE prototypes · Cityscapes-to-ACDCAvg mIoU: 61.1 -> 57.9Hurt
add entropy maximization L_max on gating values (ablation)MoE prototypes · Cityscapes-to-ACDCAvg mIoU: 61.7 -> 62.0Improved
add confidence-guided instance weighting (ablation)MoE prototypes · Cityscapes-to-ACDCAvg mIoU: 62.0 -> 62.4Improved
K-means expert generation instead of farthest point sampling or autoencoder (ablation)MoE prototypes · Cityscapes-to-ACDCmIoUImproved
use all 14 experts (K=14) instead of top-K filtering (ablation)MoE prototypes · Cityscapes-to-FogmIoUHurt
set number of experts per class N to 14 (ablation)MoE prototypes · Cityscapes-to-FogmIoUImproved
use excessively large L_max coefficient beta (ablation)MoE prototypes · Cityscapes-to-FogmIoUHurt
add MoE prototypes on SVDP baseline in continual TTASVDP · Cityscapes-to-ACDC continual TTAmean mIoU: 61.3 -> 62.5Improved
domain-shared MoE with min-max optimization vs BECoTTA domain-specific expertsBECoTTA · Cityscapes-to-ACDC continual TTAmean mIoU: 59.1 -> 62.5Improved
update model parameters during continual adaptation (TENT, DePT)TENT and DePT · Cityscapes-to-ACDC continual TTAmean mIoU: TENT 56.7 -> 55.0 over roundsHurt
method on ACDC rain vs SVDPMoE prototypes · Cityscapes-to-ACDC TTA RainmIoU: 64.2 -> 63.5Hurt

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