DermGAN: Synthetic Generation of Clinical Skin Images with Pathology

TL;DR

This paper introduces DermGAN, a GAN-based method for synthesizing realistic clinical skin images with specific conditions, aiding data augmentation and improving diagnosis of rare skin diseases.

Contribution

DermGAN adapts Pix2Pix architecture to generate diverse, high-fidelity skin images with controllable attributes, enhancing data augmentation for dermatological AI models.

Findings

Generated images are of high fidelity and visually similar to real images.

Synthetic images improve classifier performance on rare skin conditions.

DermGAN's synthetic data is effective for augmenting training datasets.

Abstract

Despite the recent success in applying supervised deep learning to medical imaging tasks, the problem of obtaining large and diverse expert-annotated datasets required for the development of high performant models remains particularly challenging. In this work, we explore the possibility of using Generative Adverserial Networks (GAN) to synthesize clinical images with skin condition. We propose DermGAN, an adaptation of the popular Pix2Pix architecture, to create synthetic images for a pre-specified skin condition while being able to vary its size, location and the underlying skin color. We demonstrate that the generated images are of high fidelity using objective GAN evaluation metrics. In a Human Turing test, we note that the synthetic images are not only visually similar to real images, but also embody the respective skin condition in dermatologists' eyes. Finally, when using the synthetic images as a data augmentation technique for training a skin condition classifier, we observe that the model performs comparably to the baseline model overall while improving on rare but malignant conditions.