Step-wise integration of deep class-specific learning for dermoscopic image segmentation

作者：

Highlights：

• A new dermoscopic segmentation method based on learning class-wise deep features.

• A class-specific model helps balance the contribution of melanoma and non-melanoma data.

• A step-wise refinement approach iteratively maximises the segmentation map agreement.

• A probability based integration approach combines the relevant complementary segmentation results.

• We achieve higher accuracy compared to the state-of-the-art methods on 3 datasets.

摘要

•A new dermoscopic segmentation method based on learning class-wise deep features.•A class-specific model helps balance the contribution of melanoma and non-melanoma data.•A step-wise refinement approach iteratively maximises the segmentation map agreement.•A probability based integration approach combines the relevant complementary segmentation results.•We achieve higher accuracy compared to the state-of-the-art methods on 3 datasets.

论文关键词：Dermoscopic,Melanoma,Segmentation,Fully convolutional networks (FCN)

论文评审过程：Received 30 November 2017, Revised 23 July 2018, Accepted 1 August 2018, Available online 2 August 2018, Version of Record 13 August 2018.

论文官网地址：https://doi.org/10.1016/j.patcog.2018.08.001