Learning Differential Diagnosis of Skin Conditions with Co-occurrence Supervision using Graph Convolutional Networks

TL;DR

This paper introduces a deep learning system that predicts differential diagnoses of skin conditions from images by leveraging co-occurrence supervision with Graph Convolutional Networks, improving accuracy over baseline models.

Contribution

The novel integration of GCNs with classification networks addresses label incompleteness and enhances diagnostic accuracy in skin condition image analysis.

Findings

Achieved 93.6% top-5 accuracy on test images.

Outperformed baseline classification network.

Effectively modeled label co-occurrence for better regularization.

Abstract

Skin conditions are reported the 4th leading cause of nonfatal disease burden worldwide. However, given the colossal spectrum of skin disorders defined clinically and shortage in dermatology expertise, diagnosing skin conditions in a timely and accurate manner remains a challenging task. Using computer vision technologies, a deep learning system has proven effective assisting clinicians in image diagnostics of radiology, ophthalmology and more. In this paper, we propose a deep learning system (DLS) that may predict differential diagnosis of skin conditions using clinical images. Our DLS formulates the differential diagnostics as a multi-label classification task over 80 conditions when only incomplete image labels are available. We tackle the label incompleteness problem by combining a classification network with a Graph Convolutional Network (GCN) that characterizes label co-occurrence and effectively regularizes it towards a sparse representation. Our approach is demonstrated on 136,462 clinical images and concludes that the classification accuracy greatly benefit from the Co-occurrence supervision. Our DLS achieves 93.6% top-5 accuracy on 12,378 test images and consistently outperform the baseline classification network.