改进U-Net的超声乳腺肿瘤分割网络

doi:10.3778/j.issn.1002-8331.2104-0189

摘要/Abstract

摘要： 乳腺超声图像具有肿瘤大小形态多变、阴影较多、边界模糊等特点，经典U-Net的乳腺肿瘤分割结果与标注图像出入较大。对此，提出改进网络MultiMixU-Net。该网络在U-Net结构中引入MultiMix block以及Respath。MultiMix block通过空洞卷积通路提高网络区分目标以及背景的能力，并通过级联该通路中各卷积层输出，融合普通卷积通路的输出来提取多尺度特征信息。Respath的改进部署使网络中收缩路径与扩张路径之间对应特征信息的传递更加有效。该改进网络在公开的超声乳腺肿瘤分割数据集上进行了测试，实验表明，MultiMixU-Net分割结果优于其他网络且参数量较少。相较于U-Net，所提网络分割结果在所有评价指标上均有提升，其中IoU、DSC分别提升0.154 1、0.127 3。

关键词: 乳腺肿瘤分割, 超声图像, U-Net, 深度学习

Abstract: The conventional U-Net performs poorly in the segmentation of breast tumors in ultrasound images, due to the variability of breast tumors’ shape, the image shadows and the obscure boundary. To overcome these difficulties, this paper proposes an enhanced network, namely MultiMixU-Net. The network introduces the MultiMix block and the Respath into the conventional U-Net. The MultiMix block uses the atrous convolution path to strengthen the ability to discern the target from the background. To extract multi-scale feature information, the MultiMix block concatenates outputs from convolution layers in the atrous convolution path and fuses outputs from the normal convolution paths. The network also modifies the layout of the Respath. This modification can make the transformation of feature information from the contract path to the expanse path more efficient. The experiments on the public ultrasound breast tumor segmentation dataset show that the MultiMixU-Net outperforms other networks and has fewer parameters. Compared with the conventional U-Net, the MultiMixU-Net in all performance metrics has better results. IoU and DSC are improved by 0.1541 and 0.1273, respectively.

Key words: breast tumor segmentation, ultrasound images, U-Net, deep learning

陈曦, 刘奇, 邓小波, 何柯辰, 全美霖. 改进U-Net的超声乳腺肿瘤分割网络[J]. 计算机工程与应用, 2022, 58(22): 219-228.

CHEN Xi, LIU Qi, DENG Xiaobo, HE Kechen, QUAN Meilin. Enhanced Network for Ultrasound Breast Tumor Segmentation Based on U-Net[J]. Computer Engineering and Applications, 2022, 58(22): 219-228.

参考文献

[1] REBECCAR L，MILLER K D，JEMAL A.Cancer statistics，2019[J].CA：A Cancer Journal for Clinicians，2019，69（1）：7-34.
[2] FAN L，STRASSER-WEIPPL K，LI J J，et al.Breast cancer in China[J].Lancet Oncology，2014，15（7）：e279-e289.
[3] HAN L Y，HUANG Y Z，DOU H R，et al.Semi-supervised segmentation of lesion from breast ultrasound images with attentional generative adversarial network[J].Computer Methods and Programs in Biomedicine，2020，189：105275.
[4] 黄永豪.基于超像素分类与多尺度注意力机制的乳腺超声图像分割算法研究[D].广州：华南理工大学，2020.
HUANG Y H.Breast ultrasound image segmentation algorithm based on super pixel classification and multi-scale attention mechanism[D].Guangzhou：South China University of Technology，2020.
[5] YAP M H，PONS G，MARTI J，et al.Automated breast ultrasound lesions detection using convolutional neural networks[J].IEEE Journal of Biomedical and Health Informatics，2017，22（4）：1218-1226.
[6] LEE H，PARK J，HWANG J Y.Channel attention module with multiscale grid average pooling for breast cancer segmentation in an ultrasound image[J].IEEE Transactions on Ultrasonics Ferroelectrics and Frequency Control，2020，67（7）：1344-1353.
[7] YEH C K，CHEN Y S，FAN W C，et al.A disk expansion segmentation method for ultrasonic breast lesions[J].Pattern Recognition，2009，42（5）：596-606.
[8] YU Y，YANG X，CHENG J，et al.Breast lesion classification based on supersonic shear-wave elastography and automated lesion segmentation from B-mode ultrasound images[J].Computers in Biology & Medicine，2017，93：31.
[9] 张新宇，刁现芬，彭韵，等.基于改进Snake模型的超声乳腺肿瘤轮廓提取[J].微计算机信息，2012，28（1）：29-31.
ZHANG X Y，DIAO X F，PENG Y，et al.Contour extraction of ultrasonic breast tumor based on modified Snake model[J].Microcomputer Information，2012，28（1）：29-31.
[10] YAP M H，GOYALA M，OSMANB F M，et al.Breast ultrasound lesions recognition：end-to-end deep learning approaches[J].Journal of Medical Imaging，2019，6（1）：011007.
[11] LONG J，SHELHAMER E，DARRELL T.Fully convolutional networks for semantic segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2017，39（4）：640-651.
[12] ZHAO H S，SHI J P，QI X J，et al.Pyramid scene parsing network[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition，2017：6230-6239.
[13] BADRINARAYANAN V，KENDALL A，CIPOLLA R.SegNet：a deep convolutional encoder-decoder architecture for image segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2017，39（12）：2481-2495.
[14] RONNEBERGER O，FISCHER P，BROX T.U-Net：convolutional networks for biomedical image segmentation[C]//18th International Conference on Medical Image Computing and Computer-Assisted Intervention.Cham：Springer，2015：234-241.
[15] CHEN L C，PAPANDREOU G，KKINOS I，et al.Semantic image segmentation with deep convolutional nets and fully connected CRFs[J].Computer Science，2014（4）：357-361.
[16] CHEN L C，PAPANDREOU G，KKINOS 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.
[17] CHEN L C，PAPANDREOU G，SCHROFF F，et al.Rethinking atrous convolution for semantic image segmentation[J].arXiv：1706.05587，2017.
[18] AMIRI M，BROOKS R，BEHBOODIB，et al.Two-stage ultrasound image segmentation using U-Net and test time augmentation[J].International Journal of Computer Assisted Radiology and Surgery，2020，15（2）：981-988.
[19] ZHUANG Z，LI N，RAJ A，et al.An RDAU-NET model for lesion segmentation in breast ultrasound images[J].PLoS One，2019，14（8）：e0221535.
[20] OKTAY O，SCHLEMPER J.Attention U-Net：learning where to look for the pancreas[J].arXiv：1804.03999，2018.
[21] IBTEHAZ N，RAHMAN M S.MultiResUNet：rethinking the U-Net architecture for multimodal biomedical image segmentation[J].Neural Networks，2019，121：74-87.
[22] IOFFE S，SZEGEDY C.Batch normalization：accelerating deep network training by reducing internal covariate shift[C]//32nd International Conference on Machine Learning，2015：448-456.
[23] PASZKE A，GROSS S，MASSA F，et al.PyTorch：an imperative style，high-performance deep learning library[C]//Advances in Neural Information Processing Systems 32，2019：8026-8037.
[24] HE K M，ZHANG X Y，REN S Q，et al.Delving deep into rectifiers：surpassing human-level performance on ImageNet classification[C]//2015 IEEE International Conference on Computer Vision，2015：1026-1034.
[25] GU Z W，CHENG J，FU H Z，et al.CE-Net：context encoder network for 2D medical image segmentation[J].IEEE Transactions on Medical Imaging，2019，38（10）：2281-2292.
[26] LOU A，GUAN S，LOEW M H，et al.DC-UNet：rethinking the U-Net architecture with dual channel efficient CNN for medical image segmentation[J].arXiv：2006.00414，2020.