Convolutional Neural Network Committees for Melanoma Classification with Classical And Expert Knowledge Based Image Transforms Data Augmentation

TL;DR

This paper explores the use of CNN committees combined with classical and expert knowledge-based data augmentation techniques to improve melanoma classification accuracy on small, unbalanced datasets, demonstrating enhanced invariance to melanoma variations.

Contribution

It introduces a novel approach of combining CNN committees with specialized data augmentation methods based on expert knowledge for melanoma detection.

Findings

Improved classification accuracy on the ISBI 2017 dataset.

Enhanced invariance to melanoma variations through expert-guided augmentation.

Effective handling of small, unbalanced datasets in medical imaging.

Abstract

Skin cancer is a major public health problem, as is the most common type of cancer and represents more than half of cancer diagnoses worldwide. Early detection influences the outcome of the disease and motivates our work. We investigate the composition of CNN committees and data augmentation for the the ISBI 2017 Melanoma Classification Challenge (named Skin Lesion Analysis towards Melanoma Detection) facing the peculiarities of dealing with such a small, unbalanced, biological database. For that, we explore committees of Convolutional Neural Networks trained over the ISBI challenge training dataset artificially augmented by both classical image processing transforms and image warping guided by specialist knowledge about the lesion axis and improve the final classifier invariance to common melanoma variations.