AI Paper Atlas

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

WOW-Seg: A Word-free Open World Segmentation Model

Danyang Li, Tianhao Wu, Bin Lin, Zhenyuan Chen, Yang Zhang, Yuxuan Li, Ming-Ming Cheng, Xiang Li

ICLR 2026 · Detection and segmentation › Open-set and few-shot detection · approach: Open-vocabulary semantic segmentation · 1 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
train with multiple masks per sample instead of single mask (ablation)WOW-Seg · LVISSemantic IoU: 74.70 -> 80.10Improved
add region decoupling of mask tokens (Cascade Attention Mask) (ablation)WOW-Seg MM · LVISSemantic IoU: 80.10 -> 82.18Improved
add output decoupling of object names (Cascade Attention Mask) (ablation)WOW-Seg MM · LVISSemantic IoU: 80.10 -> 81.94Improved
add both region and output decoupling (full Cascade Attention Mask) (ablation)WOW-Seg MM · LVISSemantic IoU: 80.10 -> 82.35Improved
add both region and output decoupling (full Cascade Attention Mask) (ablation)WOW-Seg MM · PACOSemantic IoU: 75.49 -> 79.22Improved
use Mask2Token instead of Fore2Token (ablation)WOW-Seg · LVISSemantic IoU: 72.54 -> 74.70Improved
use Mask2Token instead of Blur2Token (ablation)WOW-Seg · PACOSemantic IoU: 64.35 -> 66.04Improved
use context scale 1.5 instead of 2 in Mask2Token (ablation)WOW-Seg · LVISSemantic IoU: 82.35 -> 82.59No clear effect
increase context scale from 2 to 3.5 (ablation)WOW-Seg · LVISSemantic IoU: 82.35 -> 81.89Hurt
upgrade base VLM from InternVL2 to InternVL3 (ablation)WOW-Seg · LVISSemantic IoU: 80.8 -> 82.4Improved
upgrade base VLM from InternVL2 to InternVL3 (ablation)WOW-Seg · PACOSemantic IoU: 74.0 -> 79.2Improved
use rotated bounding-box prompts instead of ground-truth masks (ablation)WOW-Seg · LVISSemantic similarity: 89.7 -> 87.3Hurt

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