Visualizing Convolutional Neural Networks to Improve Decision Support for Skin Lesion Classification

TL;DR

This paper explores visualization techniques for CNNs in dermatology to enhance interpretability and trust in skin lesion classification, revealing that CNNs focus on features similar to dermatologists but still require further explanation.

Contribution

It demonstrates how visualizing CNN feature maps can provide insights into their decision-making process in skin lesion diagnosis, aiding medical expert understanding.

Findings

CNNs focus on dermatologically relevant features

Visualization reveals partial alignment with dermatologist reasoning

Further research needed for complete explainability

Abstract

Because of their state-of-the-art performance in computer vision, CNNs are becoming increasingly popular in a variety of fields, including medicine. However, as neural networks are black box function approximators, it is difficult, if not impossible, for a medical expert to reason about their output. This could potentially result in the expert distrusting the network when he or she does not agree with its output. In such a case, explaining why the CNN makes a certain decision becomes valuable information. In this paper, we try to open the black box of the CNN by inspecting and visualizing the learned feature maps, in the field of dermatology. We show that, to some extent, CNNs focus on features similar to those used by dermatologists to make a diagnosis. However, more research is required for fully explaining their output.