AI Paper Atlas

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

Training-Free Open-Vocabulary Camouflaged Object Segmentation via Fine-Grained Object Binding and Adaptive Hybrid Prompt

Peng Ren, Cheng Jiang, Chuande Yang, Fuming Sun, Tian Bai

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 semantic probe with LLaVA object and background descriptions (ablation)training-free OVCOS baseline · OVCamo ablation (CLIP-ViT-L/14)cIoU: 0.041 -> 0.270Improved
add SAM segmenter (ablation)training-free OVCOS · OVCamo ablation (CLIP-ViT-L/14)cIoU: 0.270 -> 0.308Improved
add adaptive hybrid prompt generation (AHPG) (ablation)training-free OVCOS · OVCamo ablation (CLIP-ViT-L/14)cIoU: 0.308 -> 0.356Improved
add entropy-guided text embedding adjustment (EGTEA) (ablation)training-free OVCOS · OVCamo ablation (CLIP-ViT-L/14)cIoU: 0.356 -> 0.371Improved
add LLaVA-generated descriptions instead of CamoTemplate prompt (ablation)semantic probe · OVCamo ablation (CLIP-ViT-L/14)cIoU: 0.333 -> 0.371Improved
use both object and background probes instead of object probe only (ablation)semantic probe · OVCamo ablation (CLIP-ViT-L/14)cIoU: 0.352 -> 0.371Improved
use both point and box prompts instead of points only (ablation)AHPG · OVCamo ablation (CLIP-ViT-L/14)cIoU: 0.346 -> 0.371Improved
use both object and background similarity maps for prompts (ablation)AHPG · OVCamo ablation (CLIP-ViT-L/14)cIoU: 0.358 -> 0.371Improved
use EGTEA instead of CASS hierarchical text embedding adjustment (ablation)text embedding adjustment · OVCamo ablation (CLIP-ViT-L/14)cIoU: 0.349 -> 0.371Improved
add background debiasing to EGTEA (ablation)EGTEA · OVCamo ablation (CLIP-ViT-L/14)cIoU: 0.354 -> 0.371Improved
use Spearman correlation instead of KL or JS divergence for patch similarity (ablation)semantic probe · OVCamo ablation (CLIP-ViT-L/14)cIoUImproved
add all components (ablation)training-free OVCOS · OVCamo ablation (CLIP-ViT-L/14)FPS: 65.1 -> 30.2Hurt

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