These are textual inversion adaptation weights for stabilityai/stable-diffusion-2-1-base trained to generate images of the epidermal nevus dermatological condition.

This model is part of the work presented in the paper Controllable Generation of Diverse Dermatological Imagery for Fair and Efficient Malignancy Classification (MICCAI 2026).

cgDDI Framework

cgDDI (Controllable Generation of Diverse Dermatological Imagery) is a hybrid framework that:

Synthesizes realistic healthy skin samples without disturbing other input properties.
Maps single-sample rare lesions onto novel skin-tones and locations non-parametrically.
Allows for efficient parametric generation with as few as 10 training samples using textual inversion and LoRA.

For more details on the dataset, training pipelines, and evaluation, please refer to the official resources:

GitHub Repository: ControllableGenDDI
Hugging Face Dataset: hcarrion/ControllableGenDDI

Citation

bibtex

@inproceedings{carrion2026cgddi,
  title     = {Controllable Generation of Diverse Dermatological Imagery for Fair and Efficient Malignancy Classification},
  author    = {Carri{\'o}n, H{\'e}ctor and Norouzi, Narges},
  booktitle = {Medical Image Computing and Computer-Assisted Intervention (MICCAI)},
  year      = {2026},
  publisher = {Springer},
  series    = {Lecture Notes in Computer Science}
}

hcarrion/epidermal_nevus

cgDDI: epidermal_nevus

cgDDI Framework

Citation

hcarrion/
epidermal_nevus