Transform domain representation-driven convolutional neural networks for skin lesion segmentation

作者：

Highlights：

• An automatic skin lesion segmentation based on deep convolution network designed.

• Image representations from transform domain used to improve the performance.

• Despite of small dataset, pre-processing and excessive data augmentation not applied.

• The analysis proved the efficiency of the proposed method on tested images.

摘要

•An automatic skin lesion segmentation based on deep convolution network designed.•Image representations from transform domain used to improve the performance.•Despite of small dataset, pre-processing and excessive data augmentation not applied.•The analysis proved the efficiency of the proposed method on tested images.

论文关键词：Convolutional neural network,Dermoscopic features,Melanoma,Skin lesion segmentation,Transform domain

论文评审过程：Received 10 August 2019, Revised 2 November 2019, Accepted 8 December 2019, Available online 9 December 2019, Version of Record 26 December 2019.

论文官网地址：https://doi.org/10.1016/j.eswa.2019.113129