Class-N-Diff: Classification-Induced Diffusion Model Can Make Fair Skin Cancer Diagnosis
This work addresses the need for better synthetic medical image generation to enhance skin cancer diagnosis, representing an incremental improvement by combining existing diffusion models with classification guidance.
The paper tackles the problem of traditional class-conditioned generative models struggling to accurately represent specific medical categories for skin cancer diagnosis by proposing Class-N-Diff, a classification-induced diffusion model that integrates a classifier to guide image generation, resulting in more realistic and diverse synthetic dermoscopic images and improved classifier performance.
Generative models, especially Diffusion Models, have demonstrated remarkable capability in generating high-quality synthetic data, including medical images. However, traditional class-conditioned generative models often struggle to generate images that accurately represent specific medical categories, limiting their usefulness for applications such as skin cancer diagnosis. To address this problem, we propose a classification-induced diffusion model, namely, Class-N-Diff, to simultaneously generate and classify dermoscopic images. Our Class-N-Diff model integrates a classifier within a diffusion model to guide image generation based on its class conditions. Thus, the model has better control over class-conditioned image synthesis, resulting in more realistic and diverse images. Additionally, the classifier demonstrates improved performance, highlighting its effectiveness for downstream diagnostic tasks. This unique integration in our Class-N-Diff makes it a robust tool for enhancing the quality and utility of diffusion model-based synthetic dermoscopic image generation. Our code is available at https://github.com/Munia03/Class-N-Diff.