An automatic retinal vessel segmentation approach based on Convolutional Neural Networks

作者：

Highlights：

• An end-to-end method based on CNN for retinal vessels segmentation;

• A new multi-encoder-decoder model is proposed taking two inputs;

• A sub-CNN is introduced to improve the extraction of vessels’ thin features;

• Improving the result through the use of dice loss function.

摘要

•An end-to-end method based on CNN for retinal vessels segmentation;•A new multi-encoder-decoder model is proposed taking two inputs;•A sub-CNN is introduced to improve the extraction of vessels’ thin features;•Improving the result through the use of dice loss function.

论文关键词：Convolutional Neural Networks,Deep learning,Semantic image segmentation,Blood vessel detection

论文评审过程：Received 25 July 2020, Revised 18 June 2021, Accepted 19 June 2021, Available online 25 June 2021, Version of Record 9 July 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115459