Bulk Production Augmentation Towards Explainable Melanoma Diagnosis

TL;DR

This paper introduces bulk production augmentation (BPA), a novel data augmentation method that generates diverse pseudo-skin tumor images to improve melanoma diagnosis and explainability, especially when limited labeled data is available.

Contribution

The paper presents BPA, a new augmentation technique that creates high-quality synthetic images with key features, enhancing melanoma detection performance over existing methods.

Findings

BPA significantly improves APN detection performance by 20 percentage points in AUC.

BPA outperforms CycleGAN-based augmentation by 11.5 to 13.7 points in AUC.

Synthetic images generated by BPA effectively boost diagnostic accuracy.

Abstract

Although highly accurate automated diagnostic techniques for melanoma have been reported, the realization of a system capable of providing diagnostic evidence based on medical indices remains an open issue because of difficulties in obtaining reliable training data. In this paper, we propose bulk production augmentation (BPA) to generate high-quality, diverse pseudo-skin tumor images with the desired structural malignant features for additional training images from a limited number of labeled images. The proposed BPA acts as an effective data augmentation in constructing the feature detector for the atypical pigment network (APN), which is a key structure in melanoma diagnosis. Experiments show that training with images generated by our BPA largely boosts the APN detection performance by 20.0 percentage points in the area under the receiver operating characteristic curve, which is 11.5 to 13.7 points higher than that of conventional CycleGAN-based augmentations in AUC.