3D-GA-Unet： MRI image segmentation algorithm for glioma based on 3D-Ghost CNN

doi:10.11772/j.issn.1001-9081.2023050606

Journal of Computer Applications ›› 2024, Vol. 44 ›› Issue (4): 1294-1302.DOI: 10.11772/j.issn.1001-9081.2023050606

• Multimedia computing and computer simulation • Previous Articles

3D-GA-Unet： MRI image segmentation algorithm for glioma based on 3D-Ghost CNN

Lijun XU, Hui LI, Zuyang LIU, Kansong CHEN(), Weixuan MA

School of Computer Science and Information Engineering，Hubei University，Wuhan Hubei 430062，China

Received:2023-05-18 Revised:2023-08-27 Accepted:2023-09-01 Online:2023-10-12 Published:2024-04-10
Contact: Kansong CHEN
About author:XU Lijun， born in 1991， Ph. D.， associate professor. Her research interests include computer vision， artificial intelligence， digital twin.
LI Hui， born in 1997， M. S. candidate. His research interests include image processing， biomedical image， deep learning.
LIU Zuyang， born in 2000， M. S. candidate. His research interests include computer vision， deep learning.
CHEN Kansong， born in 1972， Ph. D.， professor. His research interests include artificial intelligence， digital twin， industrial internet.
MA Weixuan， born in 1988， Ph. D.， lecturer. Her research interests include computer vision， remote sensing image interpretation， data mining.
Supported by:
Major Science and Technology Special Project of Hubei Province(202011901203001);Key Research and Development Program of Hubei Province(2022BAA045);Wuhan Knowledge Innovation Special Project-Shuguang Program Project(2022010801020327)

基于3D‑Ghost卷积神经网络的脑胶质瘤MRI图像分割算法3D‑GA‑Unet

许立君, 黎辉, 刘祖阳, 陈侃松(), 马为駽

湖北大学计算机与信息工程学院，武汉 430062

通讯作者: 陈侃松
作者简介:许立君（1991—），女，湖北武汉人，副教授，博士，CCF会员，主要研究方向：计算机视觉、人工智能、数字孪生
黎辉（1997—），男，湖北黄石人，主要研究方向：图像处理、生物医学图像、深度学习
刘祖阳（2000—），男，湖北武汉人，硕士研究生，主要研究方向：计算机视觉、深度学习
陈侃松（1972—），男，湖北沙市人，教授，博士生导师，博士，主要研究方向：人工智能、数字孪生、工业互联网 kschen1999@aliyun.com
马为駽（1988—），女，湖北武汉人，讲师，博士，主要研究方向：计算机视觉、遥感影像解译、数据挖掘。
基金资助:
湖北省科技重大专项(202011901203001);湖北省重点研发计划项目(2022BAA045);武汉市知识创新专项-曙光计划项目(2022010801020327)

Abstract

Abstract:

Gliomas are the most common primary cranial tumors arising from cancerous changes in the glia of the brain and spinal cord， with a high proportion of malignant gliomas and a significant mortality rate. Quantitative segmentation and grading of gliomas based on Magnetic Resonance Imaging （MRI） images is the main method for diagnosis and treatment of gliomas. To improve the segmentation accuracy and speed of glioma， a 3D-Ghost Convolutional Neural Network （CNN） -based MRI image segmentation algorithm for glioma， called 3D-GA-Unet， was proposed. 3D-GA-Unet was built based on 3D U-Net （3D U-shaped Network）. A 3D-Ghost CNN block was designed to increase the useful output and reduce the redundant features in traditional CNNs by using linear operation. Coordinate Attention （CA） block was added， which helped to obtain more image information that was favorable to the segmentation accuracy. The model was trained and validated on the publicly available glioma dataset BraTS2018. The experimental results show that 3D-GA-Unet achieves average Dice Similarity Coefficients （DSCs） of 0.863 2， 0.847 3 and 0.803 6 and average sensitivities of 0.867 6， 0.949 2 and 0.831 5 for Whole Tumor （WT）， Tumour Core （TC）， and Enhanced Tumour （ET） in glioma segmentation results. It is verified that 3D-GA-Unet can accurately segment glioma images and further improve the segmentation efficiency， which is of positive significance for the clinical diagnosis of gliomas.

Key words: glioma, medical image segmentation, neural network, attention mechanism, Convolutional Neural Network (CNN), U-Net

摘要：

脑胶质瘤是由于大脑和脊髓胶质癌变产生的、最常见的原发性颅脑肿瘤，其中恶性脑胶质瘤占比大且致死率高。利用磁共振成像（MRI）图像对脑胶质瘤定量分割和分级是目前诊治脑胶质瘤的主要方法。为提升脑胶质瘤的分割精度与速度，提出一种基于3D-Ghost卷积神经网络（CNN）的脑胶质瘤MRI图像分割算法：3D-GA-Unet。3D-GA-Unet以3D U-Net（3D U-shaped Network）为基础框架，设计基于3D-Ghost CNN模块，利用线性运算增加有用信息输出、减少传统CNN中的冗余特征；添加基于坐标注意力（CA）的模块，利于获取更多于分割精度有利的图像信息。在公共脑胶质瘤数据集BraTS2018进行训练和验证，实验结果表明，3D-GA-Unet脑胶质瘤分割结果的周围水肿区域（WT）、坏死核心区域（TC）和增强肿瘤区域（ET）的平均Dice相似系数（DSC）分别达到0.863 2、0.847 3和0.803 6，平均敏感度分别达到0.867 6、0.949 2和0.831 5。3D-GA-Unet能精准分割脑胶质瘤图像，进一步提升分割效率，对脑胶质瘤的临床诊断有积极的意义。

关键词: 脑胶质瘤, 医学图像分割, 神经网络, 注意力机制, 卷积神经网络, U-Net

CLC Number:

TP391

Lijun XU, Hui LI, Zuyang LIU, Kansong CHEN, Weixuan MA. 3D-GA-Unet： MRI image segmentation algorithm for glioma based on 3D-Ghost CNN[J]. Journal of Computer Applications, 2024, 44(4): 1294-1302.

许立君, 黎辉, 刘祖阳, 陈侃松, 马为駽. 基于3D‑Ghost卷积神经网络的脑胶质瘤MRI图像分割算法3D‑GA‑Unet[J]. 《计算机应用》唯一官方网站, 2024, 44(4): 1294-1302.

Figures/Tables 14

References 32

1	OSTROM Q T， PATIL N， CIOFFI G， et al. CBTRUS Statistical Report： primary brain and other central nervous system tumors diagnosed in the United States in 2013—2017［J］. Neuro-Oncology， 2020， 22（S1）： iv1-iv96. 10.1093/neuonc/noaa200
2	STUPP R， BRADA M， VAN DEN BENT M J， et al. High-grade malignant glioma： ESMO clinical practice guidelines for diagnosis， treatment and follow-up ［J］. Annals of Oncology， 2014， 25（3）： 93-101. 10.1093/annonc/mdu050
3	LAPOINTE S， PERRY A， BUTOWSKI N A. Primary brain tumours in adults［J］.Lancet， 2018， 392（10145）： 432-446. 10.1016/s0140-6736(18)30990-5
4	WESSELING P， CAPPER D. WHO 2016 classification of gliomas［J］. Neuropathology and Applied Neurobiology， 2018， 44（2）： 139-150. 10.1111/nan.12432
5	李洁，刘光耀，樊凤仙，等. 深度学习和影像组学在脑胶质瘤中的研究进展［J］. 磁共振成像， 2022， 13（4）：158-161. 10.12015/issn.1674-8034.2022.04.035
	LI J， LIU G Y， FAN F X， et al. Advances in deep learning and imagingomics in glioma［J］. Magnetic Resonance Imaging， 2022 13（4）： 158-161. 10.12015/issn.1674-8034.2022.04.035
6	姜迪，刘慧，李钰，等. 结合稠密特征映射的CT图像肿瘤分割模型［J］. 计算机辅助设计与图形学学报， 2021，33（8）： 1273-1286. 10.3724/sp.j.1089.2021.18656
	JIANG D， LIU H， LI Y， et al. CT image tumor segmentation model combined with dense feature mapping［J］. Journal of Computer-Aided Design & Computer Graphics， 2021， 33（8）： 1273-1286. 10.3724/sp.j.1089.2021.18656
7	CLAES A， IDEMA A J， WESSELING P. Diffuse glioma growth： a guerilla war［J］. Acta Neuropathologica， 2007， 114（5）： 443-458. 10.1007/s00401-007-0293-7
8	BELAROUSSI B， MILLES J， CARME S， et al. Intensity non-uniformity correction in MRI： existing methods and their validation［J］. Medical Image Analysis， 2006， 10（2）： 234-246. 10.1016/j.media.2005.09.004
9	FRID-ADAR M， DIAMANT I， KLANG E， et al. GAN-based synthetic medical image augmentation for increased cnn performance in liver lesion classification ［J］. Neurocomputing， 2018， 321： 321-331. 10.1016/j.neucom.2018.09.013
10	SHELHAMER E， LONG J， DARRELL T. Fully convolutional networks for semantic segmentation［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2017， 39（4）： 640-651. 10.1109/tpami.2016.2572683
11	NOH H， HONG S， HAN B. Learning deconvolution network for semantic segmentation ［C］// Proceedings of the 2015 IEEE International Conference on Computer Vision. Piscataway： IEEE， 2015：1520-1528. 10.1109/iccv.2015.178
12	CHEN C， LIU X， DING M， et al. 3D dilated multi-fiber network for real-time brain tumor segmentation in MRI［C］// Proceedings of the 2019 International Conference on Medical Image Computing and Computer-Assisted Intervention. Cham： Springer， 2019： 184-192. 10.1007/978-3-030-32248-9_21
13	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. 10.1109/tpami.2017.2699184
14	BADRINARAYANAN V， KENDALL A， CIPOLLA R. SegNet： a deep convolutional encoder-decoder architecture for image segmentation［J］. IEEE Transactions on Pattern Analysis & Machine Intelligence， 2017， 39（12）： 2481-2495. 10.1109/tpami.2016.2644615
15	DONG H， YANG G， LIU F， et al. Automatic brain tumor detection and segmentation using U-Net based fully convolutional networks［C］// Proceedings of the 2017 Annual Conference on Medical Image Understanding and Analysis. Cham： Springer， 2017： 506-517. 10.1007/978-3-319-60964-5_44
16	王海鸥，刘慧，郭强，等. 面向医学图像分割的超像素U-Net网络设计［J］. 计算机辅助设计与图形学学报， 2019， 31（6）：1007-1017. 10.3724/sp.j.1089.2019.17389
	WANG H O， LIU H， GUO Q， et al. Design of superpiexl U-Net network for medical image segmentation［J］. Journal of Computer-Aided Design & Computer Graphics， 2019，31（6）：1007-1017. 10.3724/sp.j.1089.2019.17389
17	ZHOU S， NIE D， ADELI E， et al. High-resolution encoder-decoder networks for low-contrast medical image segmentation［J］. IEEE Transactions on Image Processing， 2020， 29： 461-475. 10.1109/tip.2019.2919937
18	VALANARASU J M J， SINDAGI V A， HACIHALILOGLU I， et al. KiU-Net： overcomplete convolutional architectures for biomedical image and volumetric segmentation［J］. IEEE Transactions on Medical Imaging， 2022， 41（4）： 965-976. 10.1109/tmi.2021.3130469
19	QIN X， ZHANG Z， HUANG C， et al. U²-Net： going deeper with nested U-structure for salient object detection ［J］. Pattern Recognition， 2020， 106：107404. 10.1016/j.patcog.2020.107404
20	AHMED S F， RAHMAN F S， TABASSUM T， et al. 3D U-Net： fully convolutional neural network for automatic brain tumor segmentation［C］// Proceedings of the 2019 22nd International Conference on Computer and Information Technology. Piscataway： IEEE， 2019： 1-6. 10.1109/iccit48885.2019.9038237
21	AboELENIN N M， PIAO S， NOOR A， et al. MIRAU-Net： an improved neural network based on U-Net for gliomas segmentation［J］. Signal Processing： Image Communication， 2022， 101： 116553. 10.1016/j.image.2021.116553
22	WEN X， ZHAO B， YUAN M， et al. Application of multi-scale fusion attention U-Net to segment the thyroid gland on localized computed tomography images for radiotherapy［J］. Frontiers in Oncology， 2022， 12： 844052. 10.3389/fonc.2022.844052
23	YANG S， ZHOU X， WANG J， et al. Unsupervised domain adaptation for cross-device OCT lesion detection via learning adaptive features［C］// Proceedings of the 2020 IEEE 17th International Symposium on Biomedical Imaging. Piscataway： IEEE， 2020： 1570-1573. 10.1109/isbi45749.2020.9098380
24	PEIRIS H， HAYAT M， CHEN Z L， et al. A robust volumetric transformer for accurate 3D tumor segmentation［C］// Proceedings of the 2022 International Conference on Medical Image Computing and Computer-Assisted Intervention. Cham： Springer， 2022： 162-172. 10.1007/978-3-031-16443-9_16
25	LI G， ZHANG H， E L， et al. Recognition of honeycomb lung in CT images based on improved MobileNet model［J］. Medical Physics， 2021， 48（8）：4304-4315. 10.1002/mp.14873
26	ZHANG X， ZHOU X， LIN M， et al. ShuffleNet： an extremely efficient convolutional neural network for mobile devices［C］// Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington， DC： IEEE Computer Society， 2018：6848-6856. 10.1109/cvpr.2018.00716
27	HAN K， WANG Y， QI T， et al. GhostNet： more features from cheap operations［C］// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington， DC： IEEE Computer Society， 2020：1577-1586. 10.1109/cvpr42600.2020.00165
28	HOU Q， ZHAN D， FENG J. Coordinate attention for efficient mobile network design［C］// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington， DC： IEEE Computer Society， 2021： 13708-13717. 10.1109/cvpr46437.2021.01350
29	MENZE B H， JAKAB A， BAUER S， et al. The multimodal Brain Tumor Image Segmentation Benchmark （BRATS）［J］. IEEE Transactions on Medical Imaging， 2015， 34（10）： 1993-2024. 10.1109/tmi.2014.2377694
30	WANG G， LI W， OURSELIN S， et al. Automatic brain tumor segmentation based on cascaded convolutional neural networks with uncertainty estimation ［J］. Frontiers in Computational Neuroscience， 2019， 13： No.56. 10.3389/fncom.2019.00056
31	ZHOU Z， HE Z， JIA Y. AFPNet： a 3D fully convolutional neural network with atrous-convolution feature pyramid for brain tumor segmentation via MRI images ［J］. Neurocomputing， 2020，402： 235-244. 10.1016/j.neucom.2020.03.097
32	ZHANG D， HUANG G， ZHANG Q， et al. Exploring task structure for brain tumor segmentation from multi-modality MR images［J］. IEEE Transactions on Image Processing， 2020， 29： 9032-9043. 10.1109/tip.2020.3023609

区域名	包含真实区域	掩码值	颜色	说明
WT	周围水肿区域（ED）	2	绿色	边界包含脑胶质瘤的所有肿瘤结构
ET	增强肿瘤区域（ET）	4	黄色	通过造影剂显示的肿瘤区域
TC	非增强肿瘤核（NET）坏死核心（NCR）	1	红色	脑胶质瘤的核心部分，包含主要的肿瘤结构
背景	无	0	黑色	图像中脑胶质瘤区域外的其他区域

区域名	包含真实区域	掩码值	颜色	说明
WT	周围水肿区域（ED）	2	绿色	边界包含脑胶质瘤的所有肿瘤结构
ET	增强肿瘤区域（ET）	4	黄色	通过造影剂显示的肿瘤区域
TC	非增强肿瘤核（NET）坏死核心（NCR）	1	红色	脑胶质瘤的核心部分，包含主要的肿瘤结构
背景	无	0	黑色	图像中脑胶质瘤区域外的其他区域

模型	DSC			PPV			sensitivity			Hausdorff			参数量/10⁶	收敛时间/h
模型	WT	TC	ET	WT	TC	ET	WT	TC	ET	WT	TC	ET	参数量/10⁶	收敛时间/h
U-Net	0.838 7	0.821 5	0.768 8	0.856 1	0.853 3	0.774 5	0.857 0	0.901 4	0.829 5	2.662 3	1.751 1	2.846 6	16.323 1	5.08
GA-Unet	0.841 3	0.836 6	0.774 2	0.872 8	0.863 7	0.804 1	0.861 3	0.905 8	0.825 8	2.596 9	1.717 4	2.807 6	16.323 4	5.30
3D U-Net	0.856 2	0.851 7	0.802 3	0.885 1	0.867 7	0.815 4	0.866 9	0.914 3	0.838 6	2.373 9	1.615 3	2.704 1	16.622 9	4.94
3D-GA-Unet	0.863 2	0.847 3	0.803 6	0.867 3	0.826 9	0.983 1	0.867 6	0.949 2	0.831 5	2.079 3	1.601 2	2.573 8	16.333 0	4.50

模型	DSC			PPV			sensitivity			Hausdorff			参数量/10⁶	收敛时间/h
模型	WT	TC	ET	WT	TC	ET	WT	TC	ET	WT	TC	ET	参数量/10⁶	收敛时间/h
U-Net	0.838 7	0.821 5	0.768 8	0.856 1	0.853 3	0.774 5	0.857 0	0.901 4	0.829 5	2.662 3	1.751 1	2.846 6	16.323 1	5.08
GA-Unet	0.841 3	0.836 6	0.774 2	0.872 8	0.863 7	0.804 1	0.861 3	0.905 8	0.825 8	2.596 9	1.717 4	2.807 6	16.323 4	5.30
3D U-Net	0.856 2	0.851 7	0.802 3	0.885 1	0.867 7	0.815 4	0.866 9	0.914 3	0.838 6	2.373 9	1.615 3	2.704 1	16.622 9	4.94
3D-GA-Unet	0.863 2	0.847 3	0.803 6	0.867 3	0.826 9	0.983 1	0.867 6	0.949 2	0.831 5	2.079 3	1.601 2	2.573 8	16.333 0	4.50

模型	DSC			PPV			sensitivity			Hausdorff			参数量/10⁶	收敛迭代次数	收敛时间/h
模型	WT	TC	ET	WT	TC	ET	WT	TC	ET	WT	TC	ET	参数量/10⁶	收敛迭代次数	收敛时间/h
3D U-Net	0.856 2	0.851 7	0.802 3	0.885 1	0.867 7	0.815 4	0.866 9	0.914 3	0.838 6	2.373 9	1.615 3	2.704 1	16.322 9	84	4.94
3D U-Net+Ghost	0.857 3	0.840 3	0.793 2	0.863 4	0.868 0	0.834 1	0.862 0	0.915 8	0.843 2	2.341 2	1.605 1	2.664 2	16.322 7	82	4.92
3D U-Net+Attention	0.858 2	0.853 2	0.801 6	0.875 2	0.841 2	0.842 4	0.861 4	0.924 3	0.821 3	2.213 4	1.612 1	2.584 1	16.334 6	79	4.81
3D-GA-Unet	0.863 2	0.847 3	0.803 6	0.867 3	0.826 9	0.983 1	0.867 6	0.949 2	0.831 5	2.079 3	1.601 2	2.573 8	16.333 0	69	4.50

3D-GA-Unet： MRI image segmentation algorithm for glioma based on 3D-Ghost CNN

基于3D‑Ghost卷积神经网络的脑胶质瘤MRI图像分割算法3D‑GA‑Unet

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 14

References 32

Related Articles 15

Recommended Articles

Metrics

模型	DSC			Hausdorff
模型	WT	TC	ET	WT	TC	ET
文献［30］模型	0.873 0	0.783 0	0.754 0	5.900 0	8.030 0	4.530 0
AFPNet^［31］	0.865 8	0.768 8	0.744 3	—	—	—
Task Structure^［32］	0.896 0	0.824 0	0.782 0	—	—	—
3D-GA-Unet	0.863 2	0.847 3	0.803 6	2.079 3	1.601 2	2.573 8

[1]	Bin XIAO, Mo YANG, Min WANG, Guangyuan QIN, Huan LI. Domain generalization method of phase-frequency fusion from independent perspective [J]. Journal of Computer Applications, 2024, 44(4): 1002-1009.
[2]	Tianhua CHEN, Jiaxuan ZHU, Jie YIN. Bird recognition algorithm based on attention mechanism [J]. Journal of Computer Applications, 2024, 44(4): 1114-1120.
[3]	Bin XIAO, Yun GAN, Min WANG, Xingpeng ZHANG, Zhaoxing WANG. Network abnormal traffic detection based on port attention and convolutional block attention module [J]. Journal of Computer Applications, 2024, 44(4): 1027-1034.
[4]	Haihan WANG, Yan ZHU. Offensive speech detection with irony mechanism [J]. Journal of Computer Applications, 2024, 44(4): 1065-1071.
[5]	Quan YUAN, Changping CHEN, Ze CHEN, Linfeng ZHAN. Twice attention mechanism distantly supervised relation extraction based on BERT [J]. Journal of Computer Applications, 2024, 44(4): 1080-1085.
[6]	Rong HUANG, Junjie SONG, Shubo ZHOU, Hao LIU. Image aesthetic quality evaluation method based on self-supervised vision Transformer [J]. Journal of Computer Applications, 2024, 44(4): 1269-1276.
[7]	Jie GUO, Jiayu LIN, Zuhong LIANG, Xiaobo LUO, Haitao SUN. Recommendation method based on knowledge‑awareness and cross-level contrastive learning [J]. Journal of Computer Applications, 2024, 44(4): 1121-1127.
[8]	Pengfei ZHANG, Litao HAN, Hengjian FENG, Hongmei LI. Point cloud semantic segmentation based on attention mechanism and global feature optimization [J]. Journal of Computer Applications, 2024, 44(4): 1086-1092.
[9]	Jie WANG, Hua MENG. Image classification algorithm based on overall topological structure of point cloud [J]. Journal of Computer Applications, 2024, 44(4): 1107-1113.
[10]	Tao SUN, Zhangtian DUAN, Haonan ZHU, Peihao GUO, Heli SUN. Social event recommendation method based on unexpectedness metric [J]. Journal of Computer Applications, 2024, 44(3): 760-766.
[11]	Zhanjun JIANG, Baijing WU, Long MA, Jing LIAN. Faster-RCNN water-floating garbage recognition based on multi-scale feature and polarized self-attention [J]. Journal of Computer Applications, 2024, 44(3): 938-944.
[12]	Jingxian ZHOU, Xina LI. UAV detection and recognition based on improved convolutional neural network and radio frequency fingerprint [J]. Journal of Computer Applications, 2024, 44(3): 876-882.
[13]	Zijie HUANG, Yang OU, Degang JIANG, Cailing GUO, Bailin LI. Lightweight deep learning algorithm for weld seam surface quality detection of traction seat [J]. Journal of Computer Applications, 2024, 44(3): 983-988.
[14]	Wei SHE, Yang LI, Lihong ZHONG, Defeng KONG, Zhao TIAN. Hyperparameter optimization for neural network based on improved real coding genetic algorithm [J]. Journal of Computer Applications, 2024, 44(3): 671-676.
[15]	Yuliang ZHENG, Yunhua CHEN, Weijie BAI, Pinghua CHEN. Vehicle target detection by fusing event data and image frames [J]. Journal of Computer Applications, 2024, 44(3): 931-937.