Research map / Vision-language and multimodal / Human and embodied modeling
Align, Attend and Locate: Chest X-Ray Diagnosis via Contrast Induced Attention Network With Limited Supervision
Jingyu Liu, Gangming Zhao, Yu Fei, Ming Zhang, Yizhou Wang, Yizhou Yu
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 |
|---|---|---|---|
| add alignment module (ablation) | baseline model · NIH ChestX-ray localization, IoU 0.7 | mean localization accuracy: 0.23 -> 0.25 | Improved |
| add alignment module (ablation) | CIA-Net · NIH ChestX-ray localization, IoU 0.7 | mean localization accuracy: 0.26 -> 0.28 | Improved |
| add contrast induced attention (CIA-Net) (ablation) | baseline without alignment · NIH ChestX-ray localization, IoU 0.7 | mean localization accuracy: 0.23 -> 0.26 | Improved |
| add contrast induced attention (CIA-Net) (ablation) | baseline with alignment · NIH ChestX-ray localization, IoU 0.7 | mean localization accuracy: 0.26 -> 0.28 | Improved |
| sample negatives by structural similarity instead of random (ablation) | CIA-Net · NIH ChestX-ray localization, IoU 0.7 | mean localization accuracy: 0.27 -> 0.31 | Improved |
| sample negatives by structural similarity instead of always using canonical chest (ablation) | CIA-Net · NIH ChestX-ray localization, IoU 0.7 | mean localization accuracy: 0.21 -> 0.31 | Improved |
| increase annotated images from 40% to 80% (ablation) | CIA-Net · NIH ChestX-ray localization, IoU 0.7 | mean localization accuracy: 0.22 -> 0.31 | Improved |
| add 40% annotated images to unannotated-only training (ablation) | CIA-Net · NIH ChestX-ray localization, IoU 0.7 | mean localization accuracy: 0.13 -> 0.22 | Improved |
| use CIA-Net instead of Z, Li [14] on identification | CIA-Net · NIH ChestX-ray classification | mean AUC: 0.80 -> 0.83 | Improved |
| use CIA-Net instead of Z, Li [14] at IoU 0.5 | CIA-Net · NIH ChestX-ray localization (80% annotated + 50% unannotated) | mean localization accuracy: 0.27 -> 0.39 | Improved |
| use CIA-Net instead of Z, Li [14] at IoU 0.3 | CIA-Net · NIH ChestX-ray localization (80% annotated + 50% unannotated) | mean localization accuracy: 0.49 -> 0.53 | Improved |
| use CIA-Net instead of own base model with no annotated data at IoU 0.7 | CIA-Net · NIH ChestX-ray localization (0% annotated) | mean localization accuracy: 0.13 -> 0.13 | No clear effect |
Closest papers by meaning
- Dynamic Graph Enhanced Contrastive Learning for Chest X-Ray Report Generation (CVPR 2023 · similarity 0.86)
- Enhanced Contrastive Learning with Multi-view Longitudinal Data for Chest X-ray Report Generation (CVPR 2025 · similarity 0.86)
- CARZero: Cross-Attention Alignment for Radiology Zero-Shot Classification (CVPR 2024 · similarity 0.85)
- Bootstrapping Chest CT Image Understanding by Distilling Knowledge from X-ray Expert Models (CVPR 2024 · similarity 0.84)
- RadZero: Similarity-Based Cross-Attention for Explainable Vision-Language Alignment in Chest X-ray with Zero-Shot Multi-Task Capability (NeurIPS 2025 · similarity 0.84)
- Novel Class Discovery in Chest X-rays via Paired Images and Text (AAAI 2024 · similarity 0.83)
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.
