Collaborative Human-AI (CHAI): Evidence-Based Interpretable Melanoma Classification in Dermoscopic Images

TL;DR

This paper introduces a novel evidence-based approach for melanoma classification in dermoscopic images, combining CNN-derived embeddings with hierarchical triplet loss to improve accuracy and interpretability for both experts and non-experts.

Contribution

It proposes a new hierarchical triplet loss for joint learning of disease labels and non-expert visual similarity, enhancing interpretability and accuracy.

Findings

Improved classification accuracy over baseline models.

Enhanced relevance of localized image regions for non-experts.

Better alignment with human visual similarity metrics.

Abstract

Automated dermoscopic image analysis has witnessed rapid growth in diagnostic performance. Yet adoption faces resistance, in part, because no evidence is provided to support decisions. In this work, an approach for evidence-based classification is presented. A feature embedding is learned with CNNs, triplet-loss, and global average pooling, and used to classify via kNN search. Evidence is provided as both the discovered neighbors, as well as localized image regions most relevant to measuring distance between query and neighbors. To ensure that results are relevant in terms of both label accuracy and human visual similarity for any skill level, a novel hierarchical triplet logic is implemented to jointly learn an embedding according to disease labels and non-expert similarity. Results are improved over baselines trained on disease labels alone, as well as standard multiclass loss. Quantitative relevance of results, according to non-expert similarity, as well as localized image regions, are also significantly improved.