Research map / Vision-language and multimodal / Cross-modal learning
Cross-Modal Prototype Driven Network for Radiology Report Generation
Jun Wang, Abhir Bhalerao, Yulan He
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 cross-modal prototype network (ablation) | Transformer report generator · IU-Xray | BLEU-1: 0.467 -> 0.525 | Improved |
| add cross-modal prototype network (ablation) | Transformer report generator · MIMIC-CXR | BLEU-1: 0.321 -> 0.344 | Improved |
| add semantic prototype initialization (K-Means) instead of random (ablation) | XPRONET · IU-Xray | BLEU-4: 0.160 -> 0.199 | Improved |
| add semantic prototype initialization instead of random (ablation) | XPRONET · MIMIC-CXR | BLEU-4: 0.100 -> 0.105 | Improved |
| add improved multi-label contrastive loss (ablation) | XPRONET · IU-Xray | BLEU-4: 0.159 -> 0.199 | Improved |
| add improved multi-label contrastive loss (ablation) | XPRONET · IU-Xray | ROUGE-L: 0.377 -> 0.411 | Improved |
| add improved multi-label contrastive loss (ablation) | XPRONET · MIMIC-CXR | ROUGE-L: 0.269 -> 0.279 | Improved |
| add improved multi-label contrastive loss (ablation) | XPRONET · MIMIC-CXR | BLEU-4: 0.098 -> 0.105 | Improved |
| increase number of responding prototypes gamma from 15 to 17 (ablation) | XPRONET · IU-Xray | BLEU-4: 0.199 -> 0.171 | Hurt |
| use theta 1.750 rather than other values (ablation) | XPRONET · MIMIC-CXR | BLEU-4 | Improved |
| XPRONET versus R2GenCMN | XPRONET · IU-Xray | BLEU-4: 0.168 -> 0.199 | Improved |
| XPRONET versus R2GenCMN | XPRONET · MIMIC-CXR | BLEU-1: 0.354 -> 0.344 | Hurt |
Closest papers by meaning
- Bootstrapping Large Language Models for Radiology Report Generation (AAAI 2024 · similarity 0.88)
- Automatic Radiology Reports Generation via Memory Alignment Network (AAAI 2024 · similarity 0.88)
- S2D-Align: Shallow-to-Deep Auxiliary Learning for Anatomically-Grounded Radiology Report Generation (AAAI 2026 · similarity 0.87)
- LLM-RG4: Flexible and Factual Radiology Report Generation Across Diverse Input Contexts (AAAI 2025 · similarity 0.86)
- HC-LLM: Historical-Constrained Large Language Models for Radiology Report Generation (AAAI 2025 · similarity 0.86)
- Enhanced Contrastive Learning with Multi-view Longitudinal Data for Chest X-ray Report Generation (CVPR 2025 · 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.
