This repository contains the Textual Inversion adaptation weights trained on the hematoma disease class as part of the cgDDI framework. It was introduced in the paper Controllable Generation of Diverse Dermatological Imagery for Fair and Efficient Malignancy Classification.

These weights are learned on top of the stabilityai/stable-diffusion-2-1-base backbone to capture disease-specific dermatological concepts, allowing the controllable generation of diverse and skin-tone-balanced synthetic training datasets.

Repository: hectorcarrion/ControllableGenDDI
Dataset: hcarrion/ControllableGenDDI

Concept Token

To invoke this concept in your prompts, use the token: <hematoma-class>

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/hematoma

cgDDI: Textual Inversion for Hematoma

Concept Token

Citation

hcarrion/
hematoma