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
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use CECNet (dual-branch with COD expert) instead of fine-tuned CLIP | CLIP ViT-B/32 · CamoIT | I2T R@1: 41.3 -> 45.8 | Improved |
| use CECNet instead of fine-tuned CLIP | CLIP ViT-B/32 · CamoIT | T2I R@1: 41.1 -> 44.6 | Improved |
| add the camouflage-expert collaboration (CEC) scheme | D2S-VSE · CamoIT | I2T R@1: 37.1 -> 39.0 | Improved |
| add the camouflage-expert collaboration (CEC) scheme | AVSE · CamoIT | I2T R@1: 28.1 -> 29.9 | Improved |
| retrain on CamoIT instead of training on MS-COCO | CUSA · CamoIT | I2T R@1: 15.1 -> 23.9 | Improved |
| retrain on CamoIT instead of training on MS-COCO | D2S-VSE · CamoIT | I2T R@1: 13.3 -> 37.1 | Improved |
| add a mask-modulated input image in a single encoder (A1) (ablation) | CLIP baseline · CamoIT | I2T R@1: 41.3 -> 28.5 | Hurt |
| add mask-feature fusion in a single encoder (A2) (ablation) | CLIP baseline · CamoIT | I2T R@1: 41.3 -> 42.1 | Improved |
| add dual-branch encoder with C2GA instead of single-encoder attention (A1-A3) (ablation) | CLIP baseline · CamoIT | I2T R@1: 42.1 -> 45.8 | Improved |
| use C2GA instead of simple addition (ablation) | CECNet · CamoIT | I2T R@1: 42.5 -> 45.8 | Improved |
| use C2GA instead of linear fusion (ablation) | CECNet · CamoIT | I2T R@1: 42.4 -> 45.8 | Improved |
| use C2GA instead of vanilla graph attention (ablation) | CECNet · CamoIT | I2T R@1: 42.9 -> 45.8 | Improved |
Closest papers by meaning
- Beyond Single Images: Retrieval Self-Augmented Unsupervised Camouflaged Object Detection (ICCV 2025 · similarity 0.89)
- MM-CamObj: A Comprehensive Multimodal Dataset for Camouflaged Object Scenarios (AAAI 2025 · similarity 0.87)
- SCENIR: Visual Semantic Clarity through Unsupervised Scene Graph Retrieval (ICML 2025 · similarity 0.86)
- Instance-Level Composed Image Retrieval (NeurIPS 2025 · similarity 0.86)
- Shift the Lens: Environment-Aware Unsupervised Camouflaged Object Detection (CVPR 2025 · similarity 0.86)
- LAKE-RED: Camouflaged Images Generation by Latent Background Knowledge Retrieval-Augmented Diffusion (CVPR 2024 · similarity 0.85)
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.
