Segmentation of ischemic stroke lesion based on long-distance dependency encoding and deep residual U-Net

doi:10.11772/j.issn.1001-9081.2020111788

Abstract

Abstract: Segmenting stroke lesions automatically can provide valuable support to the clinical decision process. However, this is a challenging task due to the diversity of lesion size, shape, and location. Previous works have failed to capture global context information which is helpful to handle the diversity. To solve the problem of segmentation of ischemic stroke lesions with small sample size, an end-to-end neural network combing with residual block and non-local block on the basis of traditional U-Net was proposed to predict stroke lesion from multi-modal Magnetic Resonance Imaging (MRI) image. In this method, based on the encoder-decoder architecture of U-Net, residual blocks were stacked to solve the degradation problem and avoid the overfitting, and the non-local blocks were added to effectively encode the long-distance dependencies and provide global context information for the feature extraction process. The proposed method and its variants were evaluated on the Ischemic Stroke Lesion Segmentation (ISLES) 2017 dataset. The results showed that the proposed residual U-Net (Dice=0.29±0.23, ASSD=7.66±6.41, HD=43.71±22.11) and Residual Non-local U-Net (RN-UNet) (Dice=0.29±0.23, ASSD=7.61±6.62, HD=45.36±24.75) achieved significant improvement in all metrics compared to the baseline U-Net (Dice=0.25±0.23, ASSD=9.45±7.36, HD=54.59 ±21.19); compared with the state-of-the-art methods from ISLES website, the two methods both achieved better segmentation results, so that they can help doctors to quickly and objectively evaluate the condition of patients in clinical practices.

Key words: Acute Ischemic Stroke (AIS), residual learning, U-Net, medical image segmentation, Magnetic Resonance Imaging (MRI), deep learning

摘要： 脑卒中病灶自动分割可以为临床决策过程提供有价值的支持，而由于病灶大小、形状和位置的多样性，这项任务具有一定的挑战性。以往的研究未能很好地捕获有助于处理这种多样性的全局上下文信息。针对小样本情境下的缺血性脑卒中病灶分割这一问题，提出了在传统U-Net的基础上融合了残差模块和non-local块的端到端神经网络，用于从多模态核磁共振成像（MRI）的图像中预测卒中病灶。该方法基于U-Net的编码器-解码器结构，利用残差模块的堆叠来解决网络退化问题和避免过拟合，并通过插入non-local块编码特征图中的长距离依赖来为特征提取的过程提供全局上下文信息。对所提出的方法及其变体在缺血性卒中病灶分割挑战赛（ISLES）2017数据集上进行了评估，结果显示，所提出的残差U-Net （Dice=0.29±0.23、ASSD=7.66±6.41、HD=43.71±22.11）和RN-UNet （Dice=0.29±0.23、ASSD=7.61±6.62、HD=45.36±24.75），相对于基线U-Net （Dice=0.25±0.23、ASSD=9.45±7.36、HD=54.59±21.19）在所有指标上都有明显提升；跟ISLES官网上最先进的方法对比，所提的两个方法均取得了更好的分割结果，可见这两个方法有助于医生在临床实践中快速客观地评估病情。

关键词: 急性缺血性脑卒中, 残差学习, U-Net, 医学图像分割, 核磁共振成像, 深度学习

CLC Number:

TP391.41

HUANG Li, LU Long. Segmentation of ischemic stroke lesion based on long-distance dependency encoding and deep residual U-Net[J]. Journal of Computer Applications, 2021, 41(6): 1820-1827.

黄梨, 卢龙. 基于长距离依赖编码与深度残差U-Net的缺血性卒中病灶分割[J]. 计算机应用, 2021, 41(6): 1820-1827.

References

[1] 中国心血管健康与疾病报告编写组. 中国心血管健康与疾病报告2019概要[J]. 中国循环杂志, 2020, 35(9):833-854.(The Writing Committee of the Report on Cardiovascular Health and Diseases in China. Report on cardiovascular health and diseases in China 2019:an updated summary[J]. Chinese Circulation Journal,2020,35(9):833-854.)
[2] WANG Y,LI Z,GU H,et al. China stroke statistics 2019:a report from the national center for healthcare quality management in neurological diseases,China national clinical research center for neurological diseases, the Chinese stroke association, national center for chronic and non-communicable disease control and prevention,Chinese center for disease control and prevention and institute for global neuroscience and stroke collaborations[J]. Stroke and Vascular Neurology,2020,5(3):211-239.
[3] MATHERS C D,BOERMA T,MA FAT D. Global and regional causes of death[J]. British Medical Bulletin,2009,92:7-32.
[4] MAIER O,MENZE B H,VON DER GABLENTZ J,et al. ISLES 2015-a public evaluation benchmark for ischemic stroke lesion segmentation from multispectral MRI[J]. Medical Image Analysis, 2017,35:250-269.
[5] RONNEBERGER O,FISCHER P,BROX T. U-Net:convolutional networks for biomedical image segmentation[C]//Proceedings of the 2015 International Conference on Medical Image Computing and Computer-Assisted Intervention, LNCS 9351. Cham:Springer, 2015:234-241.
[6] ZHUANG J. LadderNet:multi-path networks based on U-Net for medical image segmentation[EB/OL].[2020-11-11]. https://arxiv.org/pdf/1810.07810.pdf.
[7] JAEGER P F,KOHL S A A,BICKELHAUPT S,et al. Retina U-Net:embarrassingly simple exploitation of segmentation supervision for medical object detection[C]//Proceedings of the 2020 Machine Learning for Health NeurIPS Workshop. New York:JMLR. org,2020:171-183.
[8] 王平, 高琛, 朱莉, 等. 基于3D深度残差网络与级联U-Net的缺血性脑卒中病灶分割算法[J]. 计算机应用, 2019, 39(11):3274-3279.(WANG P,GAO C,ZHU L,et al. Segmentation algorithm of ischemic stroke lesion based on 3D deep residual network and cascade U-Ne t[J]. Journal of Computer Applications,2019,39(11):3274-3279.)
[9] 贡荣麟, 施俊, 王骏. 面向乳腺超声图像分割的混合监督双通道反馈U-Net[J]. 中国图象图形学报, 2020, 25(10):2206-2217. (GONG R L,SHI J,WANG J. Hybrid supervised dual-channel feedback U-Net for segmentation of breast ultrasound images[J]. Journal of Image and Graphics,2020,25(10):2206-2217.)
[10] CHEN L C,PAPANDREOU G,KOKKINOS I,et al. DeepLab:semantic image segmentation with deep convolutional nets,atrous convolution,and fully connected CRFs[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40(4):834-848.
[11] YU F,KOLTUN V. Multi-scale context aggregation by dilated convolutions[EB/OL].[2020-11-11]. https://arxiv.org/pdf/1511.07122.pdf.
[12] CHEN L C,PAPANDREOU G,SCHROFF F,et al. Rethinking atrous convolution for semantic image segmentation[EB/OL].[2020-11-11]. https://arxiv.org/pdf/1706.05587.pdf.
[13] SHUAI B, ZUO Z, WANG B, et al. DAG-recurrent neural networks for scene labeling[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2016:3620-3629.
[14] PINHEIRO P O,COLLOBERT R. Recurrent convolutional neural networks for scene labeling[C]//Proceedings of the 31st International Conference on Machine Learning. New York:JMLR. org,2014:82-90.
[15] VISIN F,ROMERO A,CHO K,et al. ReSeg:a recurrent neural network-based model for semantic segmentation[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition Workshops. Piscataway:IEEE,2016:426-433.
[16] BYEON W,BREUEL T M,RAUE F,et al. Scene labeling with LSTM recurrent neural networks[C]//Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2015:3547-3555.
[17] WANG X,GIRSHICK R,GUPTA A,et al. Non-local neural networks[C]//Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE, 2018:7794-7803.
[18] WINZECK S,HAKIM A,MCKINLEY R,et al. ISLES 2016 and 2017-benchmarking ischemic stroke lesion outcome prediction based on multispectral MRI[J]. Frontiers in Neurology,2018, 9:Article No. 679.
[19] OKTAY O,SCHLEMPER J,LE FOLGOC L,et al. Attention UNet:learning where to look for the pancreas[EB/OL].[2020-11-11]. https://arxiv.org/pdf/1804.03999.pdf.
[20] LEE C-Y,XIE S,GALLAGHER P,et al. Deeply-supervised nets[C]//Proceedings of the 18th International Conference on Artificial Intelligence and Statistics. New York:JMLR. org, 2015:562-570.
[21] HE K,ZHANG X,REN S,et al. Deep residual learning for image recognition[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE, 2016:770-778.
[22] VEIT A,WILBER M J,BELONGIE S. Residual networks behave like ensembles of relatively shallow networks[C]//Proceedings of the 2016 30th International Conference on Neural Information Processing Systems. Red Hook:Curran Associates Inc.,2016:550-558
[23] HE K,ZHANG X,REN S,et al. Identity mappings in deep residual networks[C]//Proceedings of the 2016 European Conference on Computer Vision,LNCS 9908. Cham:Springer, 2016:630-645.
[24] PINTO A, MCKINLEY R, ALVES V, et al. Stroke lesion outcome prediction based on MRI imaging combined with clinical information[J]. Frontiers in Neurology, 2018, 9:Article No. 1060.
[25] JETLEY S,LORD N A,LEE N,et al. Learn to pay attention[EB/OL].[2020-11-11]. https://arxiv.org/pdf/1804.02391.pdf.
[26] ISLES. Ischemic stroke lesion segmentation challenge 2017[EB/OL].[2020-11-11]. http://www.isles-challenge.org/ISLES2017/.
[27] CHOI Y,KWON Y,PAIK M C,et al. Ischemic stroke lesion segmentation with convolutional neural networks for small data[EB/OL].[2020-11-11]. http://www.isles-challenge.org/ISLES2017/articles/choi.pdf.
[28] LUCAS C,HEINRICH M P. 2D multi-scale Res-Net for stroke segmentation[EB/OL].[2020-11-11]. http://www.isles-challenge.org/ISLES2017/articles/lucas.pdf.
[29] MOK T C W,CHUNG A C S. Deep adversarial networks for stroke lesion segmentation[EB/OL].[2020-11-11]. http://www.isles-challenge.org/ISLES2017/articles/mok.pdf.
[30] MONTEIRO M, OLIVEIRA A L. Fully convolutional neural network for 3D stroke lesion segmentation[EB/OL].[2020-11-11]. http://www.isles-challenge.org/ISLES2017/articles/monteiro.pdf.
[31] ROBBEN D,SUETENS P. Dual-scale fully convolutional neural network for final infarct prediction[EB/OL].[2020-11-11]. http://www.isles-challenge.org/ISLES2017/articles/robben.pdf.
[32] PISOV M,BELYAEV M,KRIVOV E. Neural networks ensembles for ischemic stroke lesion segmentation[EB/OL].[2020-11-11]. http://www.isles-challenge.org/ISLES2017/articles/pisov.pdf.
[33] NIU Y, GONG E, XU J, et al. Abstract WP53:improved prediction of the final infarct from acute stroke neuroimaging using deep learning[J]. Stroke,2018,49(S1):Article No. AWP53.
[34] KUMAR A,UPADHYAY N,GHOSAL P,et al. CSNet:a new DeepNet framework for ischemic stroke lesion segmentation[J]. Computer Methods and Programs in Biomedicine,2020,193:Article No. 105524.
[35] CORNELIO L K S,DEL CASTILLO M A V,NAVAL P C. UISLES:ischemic stroke lesion segmentation using U-Net[C]//Proceedings of the 2018 SAI Intelligent Systems Conference,AISC 869. Cham:Springer,2018:326-336.