一种多尺度轻量级脑胶质瘤图像分割网络

doi:10.12141/j.issn.1000-565X.220042

摘要/Abstract

摘要：

手动分割核磁共振成像（MRI）图像中的脑肿瘤区域费时、费力，容易受个人主观性的影响，能够可靠、高效的半自动或自动分割脑肿瘤，对于医学辅助诊断尤为重要。近年来，基于卷积神经网络的脑肿瘤图像自动分割方法虽然取得了长足进步，但现有方法仍未能有效地融合肿瘤图像大尺度轮廓和小尺度纹理细节等方面特征，忽略了训练时丰富的全局上下文信息。针对这些问题，文中提出了一种多尺度轻量级脑肿瘤图像分割网络MSL-Net。首先，利用改进的分层解耦卷积替换U-Net网络中的基础卷积，在高效探索多尺度多视图空间信息的同时扩大了感受野；然后，在跳跃连接处引入双向加权空洞特征金字塔结构以融合多尺度特征，并使用结合了广义Dice损失函数和Focal损失函数的混合损失函数，以提升肿瘤和非肿瘤区像素数量不平衡情况下的分割精度并加快收敛速度。在BraTS 2019数据集上的实验结果表明：文中所提出的MSL-Net网络在整体肿瘤区、核心肿瘤区和增强肿瘤区的Dice相似系数分别为0.900 3、0.830 6和0.777 0，参数量和计算量（每秒浮点运算次数）分别为3.9×10⁵和3.16×10¹⁰；与目前先进的方法相比，文中方法在实现轻量化的同时获得高的分割精度。

关键词: 脑肿瘤, 图像分割, 空洞卷积, 特征融合, 混合损失函数, 多尺度, 人工智能

Abstract:

Manual segmentation of brain tumor areas in magnetic resonance imaging (MRI) images is time-consuming and laborious, and it can be easily influenced by individual subjectivity. To reliably and efficiently segment brain tumors semi-automatically or automatically is particularly important for medically assisted diagnosis. In recent years, convolutional neural network-based methods for automatic segmentation of brain tumor images have made great progress, but the existing methods still cannot effectively fuse features in terms of large-scale contours and small-scale texture details of tumor images, and the rich global background information is ignored during trai-ning. In view of these problems, this paper proposed a multi-scale lightweight brain tumor image segmentation network MSL-Net. First, the base convolution in the U-Net network was replaced with an improved hierarchical decoupled convolution to expand the perceptual field while efficiently exploring multi-scale multi-view spatial information. Then, a bidirectional feature pyramid network structure was introduced at the skipping connection to fuse multi-scale features, and a hybrid loss function combining the generalized Dice loss function and the Focal loss function was used to improve segmentation accuracy and accelerate convergence in the case of pixel count imba-lance between tumor and non-tumor regions. Experimental results on the BraTS 2019 dataset show that the Dice similarity coefficients of the proposed MSL-Net network in the overall tumor region, core tumor region and enhanced tumor region are 0.900 3, 0.830 6 and 0.777 0, respectively, and the number of parameters and computation (floating-point operations per second) are 3.9×10⁵ and 3.16×10¹⁰, respectively. Compared with the current state-of-the-art methods, the method proposed in the paper achieves high segmentation accuracy while achieving light weight.

Key words: brain tumor, image segmentation, dilated convolution, feature fusion, hybrid loss function, multi-scale, artificial intelligence

中图分类号:

TP391

杨晋生, 陈洪鹏, 关欣, 等. 一种多尺度轻量级脑胶质瘤图像分割网络[J]. 华南理工大学学报(自然科学版), 2022, 50(12): 132-141.

YANG Jinsheng, CHEN Hongpeng, GUAN Xin, et al. A Multi-Scale Lightweight Brain Glioma Image Segmentation Network[J]. Journal of South China University of Technology(Natural Science Edition), 2022, 50(12): 132-141.

图/表 14

图1

图2

图3

图4

图5

图6

表1

表2

表3

表4

表5

表6

图7

图8

参考文献 27

1	TABATABAI G， STUPP R， van den BENT M J，et al ．Molecular diagnostics of gliomas：the clinical perspective［J］．Acta Neuropathologica，2010，120（5）：585-592.
2	WADHWA A， BHARDWAJ A， VERMA V S ．A review on brain tumor segmentation of MRI images［J］．Magnetic Resonance Imaging，2019，61：247-259.
3	CUI S， MAO L， JIANG J，et al ．Automatic semantic segmentation of brain gliomas from MRI images using a deep cascaded neural network［J］．Journal of Healthcare Engineering，2018，2018：4940593/1-15.
4	PINTO A， PEREIRA S， RASTEIRO D，et al ．Hierarchical brain tumor segmentation using extremely rando-mized trees［J］．Pattern Recognition，2018，82：105-117.
5	DEEPAK S， AMEER P M ．Brain tumor classification using deep CNN features via transfer learning［J］．Computers in Biology and Medicine，2019，111：103345/1-7.
6	TIWARI A， SRIVASTAVA S， PANT M ．Brain tumor segmentation and classification from magnetic resonance images：review of selected methods from 2014 to 2019［J］．Pattern Recognition Letters，2020，131：244-260.
7	HAVAEI M， DAVY A， WARDE-FARLEY D，et al ．Brain tumor segmentation with deep neural networks［J］．Medical Image Analysis，2017，35：18-31.
8	ZHAO X， WU Y， SONG G，et al ．A deep learning model integrating FCNNs and CRFs for brain tumor segmentation［J］．Medical Image Analysis，2018，43：98-111.
9	LONG J， SHELHAMER E， DARRELL T ．Fully convolutional networks for semantic segmentation［C］∥ Proceedings of 2015 IEEE Conference on Computer Vision and Pattern Recognition．Boston：IEEE，2015：3431-3440.
10	RONNEBERGER O， FISCHER P， BROX T ．U-Net：convolutional networks for biomedical image segmentation［C］∥ Proceedings of the 18th International Confe-rence on Medical Image Computing and Computer-Assisted Intervention．Munich：Springer，2015：234-241.
11	ÇICEK Ö， ABDULKADIR A， LIENKAMP S S，et al ．3D U-Net：learning dense volumetric segmentation from sparse annotation［C］∥ Proceedings of the 19th International Conference on Medical Image Computing and Computer-Assisted Intervention．Athens：Springer，2016：424-432.
12	HUANG C， HAN H， YAO Q，et al ．3D U²-Net：a 3D universal U-Net for multi-domain medical image segmentation［C］∥ Proceedings of the 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention．Shenzhen：Springer，2019：291-299.
13	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 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention．Shenzhen：Springer，2019：184-192.
14	DING Y， GONG L， ZHANG M，et al ．A multi-path adaptive fusion network for multimodal brain tumor segmentation［J］．Neurocomputing，2020，412：19-30.
15	HUANG G， LIU Z， van der MAATEN L，et al ．Densely connected convolutional networks［C］∥ Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition．Honolulu：IEEE，2017：4700-4708.
16	LUO Z， JIA Z， YUAN Z，et al ．HDC-Net：hierarchical decoupled convolution network for brain tumor segmentation［J］．IEEE Journal of Biomedical and Health Informatics，2020，25（3）：737-745.
17	OKTAY O， SCHLEMPER J， FOLGOC L L，et al ．Attention U-Net：learning where to look for the pancreas［EB/OL］．（2018-05-20）［2021-12-30］．.
18	IBTEHAZ N， RAHMAN M S ．MultiResUNet：rethin-king the U-Net architecture for multimodal biomedical image segmentation［J］．Neural Networks，2020，121：74-87.
19	WANG Y L， ZHAO Z J， HU S Y，et al ．CLCU-Net：cross-level connected U-shaped network with selective feature aggregation attention module for brain tumor segmentation［J］．Computer Methods and Programs in Biomedicine，2021，207：106154/1-9.
20	LI X， LUO G， WANG K ．Multi-step cascaded networks for brain tumor segmentation［C］∥ Proceedings of the 5th International Workshop on Brainlesion：Glioma，Multiple Sclerosis，Stroke and Traumatic Brain Injuries.Shenzhen：Springer，2019：163-173.
21	VALANARASU J M J， SINDAGI V A，HACIHALILOG- LU I，et al ．KiU-Net：overcomplete convolutional architectures for biomedical image and volumetric segmentation［EB/OL］．（2021-10-14）［2021-12-30］．.
22	SUDRE C H， LI W， VERCAUTEREN T，et al ．Gene-ralised dice overlap as a deep learning loss function for highly unbalanced segmentations［M］∥ Deep learning in medical image analysis and multimodal learning for clinical decision support．Berlin：Springer，2017：240-248.
23	LIN T Y， GOYAL P， GIRSHICK R，et al ．Focal loss for dense object detection［C］∥ Proceedings of 2017 IEEE International Conference on Computer Vision．Venice：IEEE，2017：2980-2988.
24	YU F， KOLTUN V， FUNKHOUSER T ．Dilated residual networks［C］∥ Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition．Honolulu：IEEE，2017：472-480.
25	MEHTA S， RASTEGARI M， CASPI A，et al ．ESPNet：efficient spatial pyramid of dilated convolutions for semantic segmentation［C］∥ Proceedings of the 15th European Conference on Computer Vision．Munich：Springer，2018：552-568.
26	ZHANG X， ZHOU X， LIN M，et al ．ShuffleNet：an extremely efficient convolutional neural network for mobile devices［C］∥ Proceedings of 2018 IEEE Conference on Computer Vision and Pattern Recognition．Salt Lake City：IEEE，2018：6848-6856.
27	TAN M， PANG R， LE Q V ．EfficientDet：scalable and efficient object detection［C］∥ Proceedings of 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition．Seattle：IEEE，2020：10781-10790.

网络结构	D_D，ET	D_D，WT	D_D，TC	D_H，ET	D_H，WT	D_H，TC
HDC	0.773 6	0.893 0	0.818 3	4.188 0	6.744 1	7.977 2
HDFU	0.771 9	0.893 8	0.818 7	3.764 4	7.359 6	7.267 3
HDFU+DHDFU	0.775 1	0.893 5	0.824 7	3.890 2	6.351 2	6.438 3
HDFU+DHDFU+DBiFPN	0.777 0	0.900 3	0.830 6	3.874 3	4.864 3	6.691 3

网络模型	D_D，ET	D_D，WT	D_D，TC
模型1	0.777 0	0.900 3	0.830 6
模型2	0.775 3	0.897 6	0.829 2
模型3	0.774 8	0.897 7	0.826 3

网络模型	D_D，ET	D_D，WT	D_D，TC
模型1	0.777 0	0.900 3	0.830 6
模型4	0.774 1	0.898 8	0.828 1
模型5	0.771 2	0.895 7	0.819 0

网络模型	D_D，ET	D_D，WT	D_D，TC
模型1	0.777 0	0.900 3	0.830 6
模型6	0.7745	0.893 7	0.826 1
模型7	0.777 5	0.898 8	0.830 2

损失函数	D_D，ET	D_D，WT	D_D，TC	D_H，ET	D_H，WT	D_H，TC
Focal 损失	0.775 9	0.898 2	0.825 9	3.872 4	5.630 7	6.576 5
广义Dice损失	0.776 9	0.899 3	0.829 0	3.930 1	4.409 3	6.814 5
混合损失	0.777 0	0.900 3	0.830 6	3.874 3	4.864 3	6.691 3