Skin cancer reorganization and classification with deep neural network

TL;DR

This paper presents a deep learning-based system for skin lesion segmentation and melanoma diagnosis, achieving high accuracy through novel neural network architectures and transfer learning techniques.

Contribution

It introduces a new segmentation neural network and a deep transfer learning model based on Inception v3 for improved melanoma detection accuracy.

Findings

High lesion boundary detection accuracy

Enhanced melanoma diagnosis performance

Effective use of transfer learning with deep neural networks

Abstract

As one kind of skin cancer, melanoma is very dangerous. Dermoscopy based early detection and recarbonization strategy is critical for melanoma therapy. However, well-trained dermatologists dominant the diagnostic accuracy. In order to solve this problem, many effort focus on developing automatic image analysis systems. Here we report a novel strategy based on deep learning technique, and achieve very high skin lesion segmentation and melanoma diagnosis accuracy: 1) we build a segmentation neural network (skin_segnn), which achieved very high lesion boundary detection accuracy; 2) We build another very deep neural network based on Google inception v3 network (skin_recnn) and its well-trained weight. The novel designed transfer learning based deep neural network skin_inceptions_v3_nn helps to achieve a high prediction accuracy.