Skin Lesion Classification Using Deep Multi-scale Convolutional Neural Networks

TL;DR

This paper introduces a multi-scale deep learning method using Inception-v3 CNNs, fine-tuned on skin lesion images, to improve classification accuracy in the ISIC 2017 challenge.

Contribution

It proposes a novel multi-scale CNN approach with transfer learning for skin lesion classification, enhancing prior methods.

Findings

Improved classification accuracy on ISIC 2017 dataset

Effective use of multi-scale inputs with Inception-v3

Demonstrated benefits of transfer learning in medical imaging

Abstract

We present a deep learning approach to the ISIC 2017 Skin Lesion Classification Challenge using a multi-scale convolutional neural network. Our approach utilizes an Inception-v3 network pre-trained on the ImageNet dataset, which is fine-tuned for skin lesion classification using two different scales of input images.