Automatic segmentation of intracerebral hemorrhage in CT images using encoder–decoder convolutional neural network
作者：
Highlights：
• We develop an encoder-decoder convolutional neural network (ED-Net).
• We introduce a synthetic loss function to overcome the data imbalanced problem.
• We propose a novel method using ED-Net for intracranial hemorrhage segmentation.
• We evaluate ED-Net on a multi-center intracranial hemorrhage clinical dataset.
• Both quantitative and visual results show ED-Net outperforms other methods.
摘要
•We develop an encoder-decoder convolutional neural network (ED-Net).•We introduce a synthetic loss function to overcome the data imbalanced problem.•We propose a novel method using ED-Net for intracranial hemorrhage segmentation.•We evaluate ED-Net on a multi-center intracranial hemorrhage clinical dataset.•Both quantitative and visual results show ED-Net outperforms other methods.
论文关键词：Intracerebral hemorrhage,Segmentation,Convolutional neural networks,Multi-scale features,Data imbalance
论文评审过程：Received 25 February 2020, Revised 24 June 2020, Accepted 1 July 2020, Available online 12 July 2020, Version of Record 12 July 2020.
论文官网地址：https://doi.org/10.1016/j.ipm.2020.102352
