Explainable skin lesion diagnosis using taxonomies

作者：

Highlights：

• A hierarchical deep neural network is designed to diagnose skin cancer.

• The proposed hierarchy is inspired by medical knowledge and improves the diagnostic performance.

• The model relies on channel and spatial attention methods to guide the network towards relevant features and regions.

• The use of attention significantly improves the explainability of the model.

摘要

•A hierarchical deep neural network is designed to diagnose skin cancer.•The proposed hierarchy is inspired by medical knowledge and improves the diagnostic performance.•The model relies on channel and spatial attention methods to guide the network towards relevant features and regions.•The use of attention significantly improves the explainability of the model.

论文关键词：Hierarchical deep learning,Explainability,Channel attention,Spatial attention,Safety-critical CADS,Skin cancer

论文评审过程：Received 15 July 2019, Revised 18 April 2020, Accepted 29 April 2020, Available online 16 May 2020, Version of Record 1 November 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107413