针对计算机辅助脑肿瘤图像边缘分割模糊、分割精度不高的问题,提出了一种改进的嵌套UNet++脑肿瘤图像分割算法.首先,设计MCAM(Mish coordinate attention module)模块代替原UNet++的特征提取部分,嵌入坐标注意力机制(coordinate attention,CA)关注不同方向上的位置信息以增强特征提取能力,使用Mish激活函数替换ReLU激活函数防止出现梯度消失,提高脑肿瘤图像分割精度和泛化能力;其次,在特征提取后加入SME(squeeze Mish excitation)模块进行挤压和激励,扩大特征图的感受野以增强对肿瘤特征的学习能力;最后,利用焦点Dice损失函数关注模糊样本的分割,从而改善脑肿瘤图像边缘分割模糊的问题.提出的算法在Figshare数据集上进行仿真实验,实验结果表明,在均值交并比(MIoU)、类别平均像素准确率(MPA)、骰子系数(Dice)和豪斯多夫距离(Hausdorff distance,HD)评估指标上分别达到83.26%、81.91%、86.45%和18.57 mm.与3DUNet、Swin-UNet、DD-UNet、LRAE-UNet和AI-UNet等算法进行对比,证明提出的算法分割效果更优.
Aiming at the problems of blurred edge segmentation and low segmentation accuracy of computer-aided brain tumor images, an improved nested UNet++ brain tumor image segmentation algorithm was proposed. Firstly, the MCAM module is designed to replace the feature extraction part of the original UNet++. The CA attention mechanism is used to focus on position information in different directions to improve feature extraction capabilities. The Mish activation function is used to replace the ReLU activation function to prevent gradient disappearance and improve segmentation. accuracy and generalization ability; then add the SME module after feature extraction for squeezing and excitation to enhance the receptive field of the feature map to enhance the learning ability of tumor features; finally, use the focus Dice loss function to focus on fuzzy samples segmentation, thereby improving the problem of blurred- DOI:10.3969/j.issn.1000-1565.2025.04.008改进UNet++模型的脑肿瘤图像分割算法付豪,张振利,陈源(江西理工大学 电气工程与自动化学院,江西 赣州 341000)摘 要:针对计算机辅助脑肿瘤图像边缘分割模糊、分割精度不高的问题,提出了一种改进的嵌套UNet++脑肿瘤图像分割算法.首先,设计MCAM(Mish coordinate attention module)模块代替原UNet++的特征提取部分,嵌入坐标注意力机制(coordinate attention,CA)关注不同方向上的位置信息以增强特征提取能力,使用Mish激活函数替换ReLU激活函数防止出现梯度消失,提高脑肿瘤图像分割精度和泛化能力;其次,在特征提取后加入SME(squeeze Mish excitation)模块进行挤压和激励,扩大特征图的感受野以增强对肿瘤特征的学习能力;最后,利用焦点Dice损失函数关注模糊样本的分割,从而改善脑肿瘤图像边缘分割模糊的问题.提出的算法在Figshare数据集上进行仿真实验,实验结果表明,在均值交并比(MIoU)、类别平均像素准确率(MPA)、骰子系数(Dice)和豪斯多夫距离(Hausdorff distance,HD)评估指标上分别达到83.26%、81.91%、86.45%和18.57 mm.与3DUNet、Swin-UNet、DD-UNet、LRAE-UNet和AI-UNet等算法进行对比,证明提出的算法分割效果更优.关键词:脑肿瘤图像分割;UNet++;MCAM;CA注意力机制;Mish激活函数;SME;焦点Dice损失函数中图分类号:TP391.41 文献标志码:A 文章编号:1000-1565(2025)04-0398-10Brain tumor image segmentation algorithm based on an improved UNet ++ modelFU Hao, ZHANG Zhenli, CHEN Yuan(School of Electrical Engineering and Automation, Jiangxi University of Science and Technology, Ganzhou 341000, China)Abstract: Aiming at the problems of blurred edge segmentation and low segmentation accuracy of computer-aided brain tumor images, an improved nested UNet++ brain tumor image segmentation algorithm was proposed. Firstly, the MCAM module is designed to replace the feature extraction part of the original UNet++. The CA attention mechanism is used to focus on position information in different directions to improve feature extraction capabilities. The Mish activation function is used to replace the ReLU activation function to prevent gradient disappearance and improve segmentation. accuracy and generalization ability; then add the SME module after feature extraction for squeezing and excitation to enhance the receptive field of the feature map to enhance the learning ability of tumor features; finally, use the focus Dice loss function to focus on fuzzy samples segmentation, thereby improving the problem of blurred- 收稿日期:2024-10-17;修回日期:2025-04-27 基金项目:国家自然科学基金项目(62363013) 第一作者:付豪(2000—),男,江西理工大学在读硕士研究生,主要从事计算机医学图像处理方向研究.E-mail:1324940506@qq.com 通信作者:张振利(1976—),男,江西理工大学副教授,主要从事人工智能、检测技术与控制方向研究.E-mail:47717770@qq.com第4期付豪等:改进UNet++模型的脑肿瘤图像分割算法河北大学学报(自然科学版) 第45卷edge segmentation of brain tumor images. The proposed algorithm was simulated on the Figshare dataset. The experimental results showed that it achieved 83.26%, 81.91%, 86.45% and 18.57mm in MIoU, MPA, Dice and Hausdorff evaluation indicators respectively. Comparison with algorithms such as 3DUnet, Swin-UNet, DD-UNet, LRAE-UNet and AI-UNet proves that the proposed algorithm has better segmentation effect.
[1] BATOOL A, BYUN Y C. Brain tumor detection with integrating traditional and computational intelligence approaches across diverse imaging modalities-Challenges and future directions[J]. Comput Biol Med, 2024, 175:108412.DOI:10.1016/j.compbiomed.2024.108412.
[2] 唐少平,陶蔷.磁共振成像在脑部疾病诊断中的研究进展[J].中国现代医生,2023,61(25): 132-135.DOI:10.3969/j.issn. 1673-9701.2023.25.031.
[3] MOHAN G, SUBASHINI MM. MRI based medical image analysis: Survey on brain tumor grade classification[J]. Biomed Signal Process Contr, 2018, 39: 139-161.DOI:10.1016/j.bspc.2017.07.007.
[4] HU H X, MAO W J, LIN Z Z, et al. Multimodal brain tumor segmentation based on an intelligent UNET-LSTM algorithm in smart hospitals[J]. ACM Trans Internet Technol, 2021, 21(3): 1-14.DOI:10.1145/3450519.
[5] SASMAL B, DHAL K G. A survey on the utilization of Superpixel image for clustering based image segmentation[J]. Multimed Tools Appl, 2023, 82(23): 35493-35555.DOI:10.1007/s11042-023-14861-9.
[6] 唐璐,赵英.基于深度学习的图像分割技术研究综述[J].电脑知识与技术,2023,19(23): 30-32.DOI:10.14004/j.cnki.ckt. 2023.1172.
[7] 王赛男,郑雄风.基于边缘计算的图像语义分割应用与研究[J].计算机科学,2020,47(Suppl.2): 276-280.DOI:10. 11896/jsjkx.200900046.
[8] 彭璟,罗浩宇,赵淦森,等.深度学习下的医学影像分割算法综述[J].计算机工程与应用,2021,57(3): 44-57.DOI:10.3778/j. issn.1002-8331.2010-0335.
[9] RONNEBERGER O, FISCHER P, BROX T.U-Net: Convolutional Networks for Biomedical Image Segmentation[M] //Medical Image Computing and Computer-Assisted Intervention - MICCAI 2015. Cham: Springer International Publishing, 2015:234-241.DOI:10.1007/978-3-319-24574-4_28.
[10] 黄敬腾,李锵,关欣.一种用于脑肿瘤分割的改进U形网络[J].工程科学学报,2023,45(6):1003-1012. DOI:10.13374/j.issn2095-9389.2022.03.25.003.
[11] CAO T Y, WANG G L, REN L L, et al.Brain tumor magnetic resonance image segmentation by a multiscale contextual attention module combined with a deep residual UNet(MCA-ResUNet)[J]. PhysMed Biol, 2022, 67(9): 095007.DOI:10.1088/1361-6560/ac5e5c.
[12] MAJI D, SIGEDAR P, SINGH M. Attention Res-UNet with guided decoder for semantic segmentation of brain tumors[J]. Biomed Signal Process Contr, 2022, 71: 103077.DOI:10.1016/j.bspc.2021.103077.
[13] 侯奕辰,彭辉,谢俊章,等.改进Unet++在脑肿瘤图像分割的研究[J].计算机工程与设计,2022,43(6):1725-1731.DOI: 10.16208/j.issn1000-7024.2022.06.029.
[14] ZHOU Z W, SIDDIQUEE MM R, TAJBAKHSH N et al. UNet++: redesigning skip connections to exploit multiscale features in image segmentation[J]. IEEE Trans on Med Imag, 2020, 39(6): 1856-1867.DOI:10.1109/TMI.2019.2959609.
[15] HU J, SHEN L, ALBANIE S, et al. Squeeze-and-excitation networks[J]. IEEE Trans Pattern Anal and Mach Intell, 2020, 42(8): 2011-2023.DOI:10.1109/TPAMI.2019.2913372.
[16] 余力, 刘宵雪,闫朝阳,等.面向多模态MRI脑胶质瘤区域三维分割与生存期预测的级联U-Net网络[J].中国图象图形学报,2022,27(3): 850-861.DOI:10.11834/jig.210137.
[17] 张晓倩,罗建,杨梅,等.基于改进U-Net模型和CBAM的脑肿瘤MRI图像分割[J].现代计算机,2022,28(16): 60-66.DOI:10.3969/j.issn.1007-1423.2022.16.012.
[18] 郑伟,赵金芳,张奕婧,等.基于感受野扩增和注意力机制的U-Net脑肿瘤MR图像分割[J].河北大学学报(自然科学版),2022,42(5): 542-551.DOI:10.3969/j.issn.1000-1565.2022.05.013.
[19] HOU Q B, ZHOU D Q, FENG J S. Coordinate attention for efficient mobile network design[EB/OL].2021:2103.02907. https://arxiv.org/abs/2103.02907v1.
[20] MISRA D. Mish:A self regularized non-monotonic neural activation function[J].Statistics,2019.DOI:10.48550/arXiv. 1908.08681.
[21] JADON S. A survey of loss functions for semantic segmentation[C] //2020 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology(CIBCB), Via del Mar, Chile, IEEE, 2020: 1-7.DOI:10.1109/CIBCB48159.2020.9277638.
[22] 袁雪,陈茂洲.基于3D-Unet的sMRI图像分割研究[J].现代电子技术,2023,46(23): 71-74.DOI:10.16652/j.issn.1004-373x. 2023.23.013.
[23] LIAO Z Q, PENG H, LIU T. Brain tumor segmentation based on improved swin-UNet[C] // 2023 2nd International Conference on Artificial Intelligence and Intelligent Information Processing(AIIIP). Hangzhou, China, IEEE, 2023:222-225.DOI:10.1109/AIIIP61647. 2023.00048.
[24] 苏赋,方东,王龙业,等.基于双解码路径DD-UNet的脑肿瘤图像分割算法[J].光电子·激光,2023,34(3): 328-336.DOI:10.16136/j.joel.2023.03.0203.
[25] 林嘉豪,王瑜,肖洪兵,等.LRAE-Unet:轻量级MRI脑肿瘤全自动分割网络[J].中国医学物理学杂志,2024,41(1):43-49.DOI:10.3969/j.issn.1005-202X.2024.01.006.
[26] 黄炜庭.基于AI UNet的胶质瘤病变图像分割[D].兰州:兰州大学,2024.DOI:10.27204/d.cnki.glzhu.2023.001535. (
