# LesionGen: A Concept-Guided Diffusion Model for Dermatology Image Synthesis

arXiv: https://arxiv.org/abs/2507.23001 · HTML: https://mlsanity.com/paper/2507.23001/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Medical imaging and segmentation, one of 14 directions in Diffusion models.
- Builds on 13 of its 25 references that are in the map.
- First posted on arXiv 30 Jul 2025.

## Placed in

- Diffusion models / [Medical imaging and segmentation](https://mlsanity.com/diffusion/medical-imaging-and-segmentation/)

## Builds on (most cited first, 13 of 13)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [LAION-5B: An open large-scale dataset for training next generation image-text models](https://mlsanity.com/paper/2210.08402/) (2022)
- [ChestX-ray8: Hospital-scale Chest X-ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases](https://mlsanity.com/paper/1705.02315/) (2017)
- [CoLa-Diff: Conditional Latent Diffusion Model for Multi-Modal MRI Synthesis](https://mlsanity.com/paper/2303.14081/) (2023)
- [PathLDM: Text conditioned Latent Diffusion Model for Histopathology](https://mlsanity.com/paper/2309.00748/) (2023)
- [Diffusion-based Data Augmentation for Skin Disease Classification: Impact Across Original Medical Datasets to Fully Synthetic Images](https://mlsanity.com/paper/2301.04802/) (2023)
- [Augmenting medical image classifiers with synthetic data from latent diffusion models](https://mlsanity.com/paper/2308.12453/) (2023)
- [Medical diffusion on a budget: Textual Inversion for medical image generation](https://mlsanity.com/paper/2303.13430/) (2023)
- [BiasPruner: Debiased Continual Learning for Medical Image Classification](https://mlsanity.com/paper/2407.08609/) (2024)
- [Empirical Validation of Conformal Prediction for Trustworthy Skin Lesions Classification](https://mlsanity.com/paper/2312.07460/) (2023)
- [EyeDiff: text-to-image diffusion model improves rare eye disease diagnosis](https://mlsanity.com/paper/2411.10004/) (2024)
- [Foundation Models as Class-Incremental Learners for Dermatological Image Classification](https://mlsanity.com/paper/2507.14050/) (2025)

## Built on by (most cited first, 4 of 4)

- [CRAFT: Clinical Reward-Aligned Finetuning for Medical Image Synthesis](https://mlsanity.com/paper/2605.12650/) (2026)
- [CSEval: A Framework for Evaluating Clinical Semantics in Text-to-Image Generation](https://mlsanity.com/paper/2602.12004/) (2026)
- [DermaFlux: Synthetic Skin Lesion Generation with Rectified Flows for Enhanced Image Classification](https://mlsanity.com/paper/2603.16392/) (2026)
- [What Makes a Good Medical Image Tokenizer? Rethinking Reconstruction and Generation in Medical Image Tokenization](https://mlsanity.com/paper/2609.24691/) (2026)

## Abstract (arXiv)

Deep learning models for skin disease classification require large, diverse, and well-annotated datasets. However, such resources are often limited due to privacy concerns, high annotation costs, and insufficient demographic representation. While text-to-image diffusion probabilistic models (T2I-DPMs) offer promise for medical data synthesis, their use in dermatology remains underexplored, largely due to the scarcity of rich textual descriptions in existing skin image datasets. In this work, we introduce LesionGen, a clinically informed T2I-DPM framework for dermatology image synthesis. Unlike prior methods that rely on simplistic disease labels, LesionGen is trained on structured, concept-rich dermatological captions derived from expert annotations and pseudo-generated, concept-guided reports. By fine-tuning a pretrained diffusion model on these high-quality image-caption pairs, we enable the generation of realistic and diverse skin lesion images conditioned on meaningful dermatological descriptions. Our results demonstrate that models trained solely on our synthetic dataset achieve classification accuracy comparable to those trained on real images, with notable gains in worst-case subgroup performance. Code and data are available here.