Research map / Image generation / Autoregressive and latent generation
Aligning Synthetic Medical Images with Clinical Knowledge using Human Feedback
Shenghuan Sun, Gregory Goldgof, Atul Butte, Ahmed Alaa
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 |
|---|---|---|---|
| finetune the diffusion model with pathologist feedback via reward model | conditional diffusion model · clinical plausibility of 400 synthetic images | fraction plausible: 0.21 to 0.75 | Improved |
| finetune with pathologist feedback vs no feedback | conditional diffusion model · Precision, Recall, Coverage | fidelity and diversity | Improved |
| use automatic classifier-guidance feedback instead of pathologist feedback (ablation) | conditional diffusion model · cell classification with synthetic training data | accuracy | No clear effect |
| add real data augmentation to the reward model training (D_s + D_r) (ablation) | reward model · cell classification with synthetic training data | accuracy | Improved |
| train classifier on synthetic data with pathologist feedback vs without | ResNeXt-50 classifier · 16-class cell classification on real test data | accuracy | Improved |
| increase pathologist-labeled images from 0% to 10%, 50%, 100% of 3,936 (ablation) | reward model · cell classification with synthetic training data | accuracy | Improved |
| replace conditional diffusion with conditional GAN as base generator (ablation) | generative model · cell classification on real test data | AUC: 0.929482 -> 0.413551 | Hurt |
| replace conditional diffusion with latent diffusion as base generator (ablation) | generative model · cell classification on real test data | AUC: 0.929482 -> 0.566388 | Hurt |
| add band and segmented Neutrophil subtype annotations to the reward | finetuned diffusion model · Neutrophil subtype classification | accuracy: 71.88 | Improved |
Closest papers by meaning
- Doctor Approved: Generating Medically Accurate Skin Disease Images through AI-Expert Feedback (NeurIPS 2025 · similarity 0.89)
- Medical thinking with multiple images (ICLR 2026 · similarity 0.87)
- How Do Medical MLLMs Fail? A Study on Visual Grounding in Medical Images (ICLR 2026 · similarity 0.86)
- Breaking the Barrier of Hard Samples: A Data-Centric Approach to Synthetic Data for Medical Tasks (ICML 2025 · similarity 0.86)
- Trade-Offs in Fine-Tuned Diffusion Models between Accuracy and Interpretability (AAAI 2024 · similarity 0.86)
- Tokenizer-Generator Coupling in Medical Image Generation (NeurIPS 2026 · 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.
