AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Camouflage-aware Image-Text Retrieval via Expert Collaboration

Yao Jiang, Zhongkuan Mao, Xuan Wu, Keren Fu, Qijun Zhao

CVPR 2026 · Vision-language and multimodal › Cross-modal learning · approach: Probabilistic embeddings for cross-modal retrieval · 0 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
use CECNet (dual-branch with COD expert) instead of fine-tuned CLIPCLIP ViT-B/32 · CamoITI2T R@1: 41.3 -> 45.8Improved
use CECNet instead of fine-tuned CLIPCLIP ViT-B/32 · CamoITT2I R@1: 41.1 -> 44.6Improved
add the camouflage-expert collaboration (CEC) schemeD2S-VSE · CamoITI2T R@1: 37.1 -> 39.0Improved
add the camouflage-expert collaboration (CEC) schemeAVSE · CamoITI2T R@1: 28.1 -> 29.9Improved
retrain on CamoIT instead of training on MS-COCOCUSA · CamoITI2T R@1: 15.1 -> 23.9Improved
retrain on CamoIT instead of training on MS-COCOD2S-VSE · CamoITI2T R@1: 13.3 -> 37.1Improved
add a mask-modulated input image in a single encoder (A1) (ablation)CLIP baseline · CamoITI2T R@1: 41.3 -> 28.5Hurt
add mask-feature fusion in a single encoder (A2) (ablation)CLIP baseline · CamoITI2T R@1: 41.3 -> 42.1Improved
add dual-branch encoder with C2GA instead of single-encoder attention (A1-A3) (ablation)CLIP baseline · CamoITI2T R@1: 42.1 -> 45.8Improved
use C2GA instead of simple addition (ablation)CECNet · CamoITI2T R@1: 42.5 -> 45.8Improved
use C2GA instead of linear fusion (ablation)CECNet · CamoITI2T R@1: 42.4 -> 45.8Improved
use C2GA instead of vanilla graph attention (ablation)CECNet · CamoITI2T R@1: 42.9 -> 45.8Improved

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