AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

CDICS: Delving Into Fine-Grained Attribute for In-Context Segmentation via Compositional Prompts and Phased Decoupling

Zhiyu Li, Dianmo Sheng, Qi Chu, Shilong Chen, Tao Gong, Zhou Wei, Nenghai Yu

CVPR 2026 · Detection and segmentation › Open-set and few-shot detection · approach: Open-vocabulary semantic segmentation

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 Appearance Fusion module to the baseline (ablation)CDICS baseline · ColorPACOIoU: 50.3 -> 51.4Improved
add Appearance Fusion module to the baseline (ablation)CDICS baseline · ColorPACOFPR: 12.0 -> 9.3Improved
add decoupled two-stage training on top of Appearance Fusion (ablation)CDICS · ColorPACOIoU: 51.4 -> 58.9Improved
add decoupled two-stage training on top of Appearance Fusion (ablation)CDICS · ColorPACOFPR: 9.3 -> 5.7Improved
add decoupled two-stage training on top of Appearance Fusion (ablation)CDICS · ColorPACOAP50: 30.8 -> 42.4Improved
add compositional prompts and phased decoupling to SINESINE · ColorPACOIoU: 44.1 -> 57.6Improved
add compositional prompts and phased decoupling to SINESINE · ColorPACOFPR: 17.1 -> 3.9Improved
add compositional prompts and phased decoupling to SINESINE · COCO-20iIoU: 64.5 -> 65.2No clear effect
add compositional prompts and phased decoupling to SINESINE · ColorPartImageNet (out-of-domain)FPR: 23.4 -> 12.5Improved
add compositional prompts and phased decoupling to SINESINE · ColorPACO instance segmentationAP: 14.3 -> 20.8Improved
compare against SegGPT on out-of-domain IoUCDICS · ColorPartImageNet (out-of-domain)IoU: 85.2 (SegGPT) vs 84.4 (CDICS)Hurt

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