基于精细化残差U-Net的新生儿局灶性脑白质损伤分割模型

doi:10.11772/j.issn.1001-9081.2019049101

计算机应用 ›› 2019, Vol. 39 ›› Issue (12): 3456-3461.DOI: 10.11772/j.issn.1001-9081.2019049101

• 第十七届中国机器学习会议(CCML 2019)论文 • 上一篇下一篇

基于精细化残差U-Net的新生儿局灶性脑白质损伤分割模型

刘亚龙, 李洁, 王颖, 仵赛飞, 邹佩

西安电子科技大学电子工程学院, 西安 710071

收稿日期:2019-04-29 修回日期:2019-06-26 出版日期:2019-12-10 发布日期:2019-12-17
作者简介:刘亚龙(1994-),男,江苏苏州人,硕士研究生,CCF会员,主要研究方向:深度学习、医学图像分割;李洁(1972-),女,陕西西安人,教授,博士,主要研究方向:图像处理、智慧交通;王颖(1981-),女,陕西西安人,副教授,博士,CCF会员,主要研究方向:影像处理与分析、模式识别;仵赛飞(1995-),男,河南省周口人,硕士研究生,主要研究方向:机器学习、计算机视觉;邹佩(1993-),女,湖北孝感人,硕士研究生,主要研究方向:影像处理与分析。
基金资助:
国家自然科学基金资助项目（61671339）。

Segmentation model of neonatal punctate white matter lesion based on refined deep residual U-Net

LIU Yalong, LI Jie, WANG Ying, WU Saifei, ZOU Pei

School of Electronic Engineering, Xidian University, Xi'an Shaanxi 710071, China

Received:2019-04-29 Revised:2019-06-26 Online:2019-12-10 Published:2019-12-17
Contact: 刘亚龙
Supported by:
This work is partially supported by the National Natural Science Foundation of China (61671339).

摘要/Abstract

摘要： 针对新生儿局灶性脑白质损伤的病灶区域小而样本差异大导致的检测与分割病灶较为困难的问题，提出一种精细化深度残差U-Net模型，以对病灶进行精细的语义分割。首先，把核磁共振（MRI）图像裁剪成较小的图像块；其次，利用残差U-Net提取出每个图像块不同层次的深度特征；然后，将特征进行融合并输出每个图像块的病灶分布概率图；最后，由全连接条件随机场对拼接后的概率图进行优化得到最终的分割结果。在某合作医院提供的数据集上的评估结果显示，在仅使用T1序列单模态数据的情况下，该模型在分割新生儿局灶性脑白质损伤时，病灶边缘的分割精度得到提高，且模型抗干扰能力较好。该模型的Dice相似性系数达到了62.51%，敏感度达到69.76%，特异性达到99.96%，修正的Hausdorff距离降低到33.67。

关键词: 局灶性脑白质损伤, 新生儿, 脑部肿瘤分割, 深度学习, 语义分割, 深度残差U-Net模型

Abstract: The tiny lesion area and the large difference between samples of neonatal punctate white matter lesion make it difficult to detect and segment the lesion. To solve the problem, a refined deep residual U-Net was proposed to realize the fine semantic segment of the lesion. Firstly, a Magnetic Resonance Imaging (MRI) image was cut into small patches. Secondly, the deep features of multiple layers of each image patch were extracted by the residual U-Net. Then, the features were fused and the probability map of the lesion distribution of each image patch was obtained. Finally, the probability map after splicing was optimized by the fully-connected condition random field to obtain the final segmentation results. The performance of the algorithm was evaluated on a dataset provided by a cooperative hospital. The results show that with only T1 order unimodal data used, the proposed model has the lesion's edge segmented more precisely, and the anti-interference ability of the model is prominent. The model has the Dice similarity coefficient of 62.51%, the sensitivity of 69.76%, the specificity of 99.96%, and the modified Hausdorff distance reduced to 33.67.

Key words: Punctate White Matter Lesion (PWML), neonate, brain tumor segmentation, deep learning, semantic segmentation, deep residual U-Net

中图分类号:

TP391.4

刘亚龙, 李洁, 王颖, 仵赛飞, 邹佩. 基于精细化残差U-Net的新生儿局灶性脑白质损伤分割模型[J]. 计算机应用, 2019, 39(12): 3456-3461.

LIU Yalong, LI Jie, WANG Ying, WU Saifei, ZOU Pei. Segmentation model of neonatal punctate white matter lesion based on refined deep residual U-Net[J]. Journal of Computer Applications, 2019, 39(12): 3456-3461.

参考文献

[1] TORTORA D, PANARA V, MATTEI P A, et al. Comparing 3T T1-weighted sequences in identifying hyperintense punctate lesions in preterm neonates[J]. American Journal of Neuroradiology, 2015, 36(3):581-586.
[2] KERSBERGEN K J, BENDERS M J N L, GROENENDAAL F, et al. Different patterns of punctate white matter lesions in serially scanned preterm infants[J]. PloS One, 2014, 9(10):e108904.
[3] DYET L E, KENNEA N, COUNSELL S J, et al. Natural history of brain lesions in extremely preterm infants studied with serial magnetic resonance imaging from birth and neurodevelopmental assessment[J]. Pediatrics, 2006, 118(2):536-548.
[4] CORNETTE L G, TANNER S F, RAMENGHI L A, et al. Magnetic resonance imaging of the infant brain:anatomical characteristics and clinical significance of punctate lesions[J]. Archives of Disease in Childhood-Fetal and Neonatal Edition, 2002, 86(3):F171-F177.
[5] CHENG I, HAJARI N, FIROUZMANESH A, et al. White matter injury detection in neonatal MRI[C]//Proceedings of the Medical Imaging 2013:Computer-Aided Diagnosis, SPIE 8670. Bellingham, WA:SPIE, 2013:86702L.
[6] CHENG I, MILLER S P, DUERDEN E G, et al. Stochastic process for white matter injury detection in preterm neonates[J]. NeuroImage:Clinical, 2015, 7:622-630.
[7] MUKHERJEE S, CHENG I, MILLER S, et al. A fast segmentation-free fully automated approach to white matter injury detection in preterm infants[J]. Medical and Biological Engineering and Computing, 2019, 57(1):71-87.
[8] MATAS J, CHUM O, URBAN M, et al. Robust wide-baseline stereo from maximally stable extremal regions[J]. Image and Vision Computing, 2004, 22(10):761-767.
[9] JIAO Z, GAO X, WANG Y, et al. A deep feature based framework for breast masses classification[J]. Neurocomputing, 2016, 197:221-231.
[10] HU Y, LI J, JIAO Z. Mammographic mass detection based on saliency with deep features[C]//Proceedings of the 2016 International Conference on Internet Multimedia Computing and Service. New York:ACM, 2016:292-297.
[11] JIAO Z, GAO X, WANG Y, et al. A parasitic metric learning net for breast mass classification based on mammography[J]. Pattern Recognition, 2018, 75:292-301.
[12] YANG D, WANG Y, JIAO Z. Asymmetry analysis with sparse autoencoder in mammography[C]//Proceedings of the 2016 International Conference on Internet Multimedia Computing and Service. New York:ACM, 2016:287-291.
[13] JIAO Z, GAO X, WANG Y. Deep convolutional neural networks for mental load classification based on EEG data[J]. Pattern Recognition, 2018, 76:582-595.
[14] 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.
[15] HAVAEI M, DAVY A, WARDE-FARLEYC D. Brain tumor segmentation with deep neural networks[J]. Medical Image Analysis, 2017, 35:18-31.
[16] GUERRERO R, QIN C, OKTAY O, et al. White matter hyperintensity and stroke lesion segmentation and differentiation using convolutional neural networks[J]. NeuroImage:Clinical, 2018, 17:918-934.
[17] 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.
[18] ZHANG Z, LIU Q, WANG Y. Road extraction by deep residual U-Net[J]. IEEE Geoscience and Remote Sensing Letters, 2018, 15(5):749-753.
[19] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway:IEEE, 2017:2999-3007.
[20] KRÄHENBÜHL P, KOLTUN V. Efficient inference in fully connected CRFs with Gaussian edge potentials[C]//Proceedings of the 25th Annual Conference on Neural Information Processing Systems. New York:Curran Associates, 2011:109-117.
[21] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway:IEEE, 2017:2980-2988.
[22] LAFFERTY J, MCCALLUM A, PEREIRA F C N. Conditional random fields:probabilistic models for segmenting and labeling sequence data[C]//Proceedings of the 18th International Conference on Machine Learning. New York:International Machine Learning Society, 2001:282-289.
[23] ZHENG S, JAYASUMANA S, ROMERA-PAREDES B, et al. Conditional random fields as recurrent neural networks[C]//Proceedings of the 2015 IEEE International Conference on Computer Vision. Piscataway:IEEE, 2015:1529-1537.
[24] KINGMA D P, BA J L. Adam:a method for stochastic optimization[EB/OL].[2019-02-20]. http://de.arxiv.org/pdf/1412.6980.
[25] SNOEK J, LAROCHELLE H, ADAMS R P. Practical Bayesian optimization of machine learning algorithms[C]//Proceedings of the 26th Annual Conference on Neural Information Processing Systems. New York:Curran Associates, 2012:2951-2959.
[26] HE K, GKIOXARI G, DOLÁR P, et al. Mask R-CNN[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway:IEEE, 2017:2980-2988.

基于精细化残差U-Net的新生儿局灶性脑白质损伤分割模型

Segmentation model of neonatal punctate white matter lesion based on refined deep residual U-Net

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	陈成瑞, 孙宁, 何世彪, 廖勇. 面向C-V2X通信的基于深度学习的联合信道估计与均衡算法[J]. 计算机应用, 2021, 41(9): 2687-2693.
[2]	郑志强, 胡鑫, 翁智, 王雨禾, 程曦. 基于改进DenseNet的牛眼图像特征提取方法[J]. 计算机应用, 2021, 41(9): 2780-2784.
[3]	谢德峰, 吉建民. 融入句法感知表示进行句法增强的语义解析[J]. 计算机应用, 2021, 41(9): 2489-2495.
[4]	代雨柔, 杨庆, 张凤荔, 周帆. 基于自监督学习的社交网络用户轨迹预测模型[J]. 计算机应用, 2021, 41(9): 2545-2551.
[5]	赵宏, 孔东一. 图像特征注意力与自适应注意力融合的图像内容中文描述[J]. 计算机应用, 2021, 41(9): 2496-2503.
[6]	徐江浪, 李林燕, 万新军, 胡伏原. 结合目标检测的室内场景识别方法[J]. 计算机应用, 2021, 41(9): 2720-2725.
[7]	何正海, 线岩团, 王蒙, 余正涛. 融合句法指导与字符注意力机制的案情阅读理解方法[J]. 计算机应用, 2021, 41(8): 2427-2431.
[8]	曹玉红, 徐海, 刘荪傲, 王紫霄, 李宏亮. 基于深度学习的医学影像分割研究综述[J]. 计算机应用, 2021, 41(8): 2273-2287.
[9]	秦斌斌, 彭良康, 卢向明, 钱江波. 司机分心驾驶检测研究进展[J]. 计算机应用, 2021, 41(8): 2330-2337.
[10]	杜炎, 吕良福, 焦一辰. 基于模糊推理的模糊原型网络[J]. 计算机应用, 2021, 41(7): 1885-1890.
[11]	侯笑晗, 金国栋, 谭力宁, 薛远亮. 基于自适应和最优特征的合成孔径雷达舰船检测方法[J]. 计算机应用, 2021, 41(7): 2150-2155.
[12]	李亚芳, 梁烨, 冯韦玮, 祖宝开, 康玉健. 基于社区优化的深度网络嵌入方法[J]. 计算机应用, 2021, 41(7): 1956-1963.
[13]	高钦泉, 黄炳城, 刘文哲, 童同. 基于改进CenterNet的竹条表面缺陷检测方法[J]. 计算机应用, 2021, 41(7): 1933-1938.
[14]	冯兴杰, 张天泽. 基于分组卷积进行特征融合的全景分割算法[J]. 计算机应用, 2021, 41(7): 2054-2061.
[15]	王月, 江逸茗, 兰巨龙. 基于改进三元组网络和K近邻算法的入侵检测[J]. 计算机应用, 2021, 41(7): 1996-2002.