AI Paper Atlas

Research map / Detection and segmentation / CNN features

ReCLIP++: Learn to Rectify the Bias of CLIP for Unsupervised Semantic Segmentation

Jingyun Wang, Guoliang Kang

IJCV 2025 · Detection and segmentation › CNN features · approach: Context modeling for semantic segmentation · 8 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 decoder (ablation)ReCLIP (rec.) · PASCAL VOCmIoU: 58.5 -> 75.6Improved
add label hypothesis (ablation)ReCLIP (rec.) + Decoder · PASCAL VOCmIoU: 75.6 -> 77.3Improved
add independent bias encoding (addition fusion) (ablation)ReCLIP++ · PASCAL VOCmIoU: 77.3 -> 81.1Improved
fuse W_r and W_p by matrix multiplication instead of addition (ablation)ReCLIP++ · PASCAL VOCmIoU: 81.1 -> 85.4Improved
fix Query instead of making it learnable (ablation)ReCLIP++ · PASCAL VOCmIoU: 11.6 -> 85.4Improved
use element-wise subtraction instead of addition (ablation)ReCLIP++ · PASCAL VOCmIoU: 77.7 -> 85.4Improved
freeze CLIP instead of fine-tuning its image encoder (ablation)ReCLIP++ · PASCAL VOCmIoU: 76.5 -> 85.4Improved
rectify both class and space preference instead of class only (ablation)ReCLIP++ · PASCAL VOCmIoU: 76.4 -> 85.4Improved
rectify both class and space preference instead of space only (ablation)ReCLIP++ · PASCAL ContextmIoU: 32.4 -> 36.1Improved
add Z as decoder input (ablation)ReCLIP++ · PASCAL VOCmIoU: 73.6 -> 85.4Improved
add distillation stage (ablation)ReCLIP++ · PASCAL VOCmIoU: 85.4 -> 80.0Hurt
add distillation stage (ablation)ReCLIP++ · ADE20KmIoU: 16.4 -> 15.8Hurt

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