YANG Mei,LUO Jian,ZHANG Xiaoqian,et al.A Novel Coronary Pneumonia Image Segmentation Method based on Improved U-Net[J].Journal of Chengdu University of Information Technology,2023,38(01):44-48.[doi:10.16836/j.cnki.jcuit.2023.01.007]
基于改进U-Net的新冠肺炎图像分割方法
- Title:
- A Novel Coronary Pneumonia Image Segmentation Method based on Improved U-Net
- 文章编号:
- 2096-1618(2023)01-0044-05
- Keywords:
- U-Net; Gaussian error linear element; channel attention; spatial attention; mixed loss function
- 分类号:
- TP391.41
- 文献标志码:
- A
- 摘要:
- 近两年,新冠肺炎在全球暴发,给人类带来了极其严重的生命安全隐患,为尽可能地提高医生的诊断效率,研究新冠肺炎图像的病灶分割方法是极具价值的。利用U-Net网络模型作为基础网络,在第一、二层拼接前,引入改进的通道注意力,强化重要信息; 在第三、四层拼接前,引入空间注意力,提取空间信息,对目标区域的细节信息进行补充; 最后使用混合损失函数,加快网络收敛速度,避免样本不均衡。对比基础网络,改进后的网络模型既能够分割出较大的目标区域,也能够分割出较小的目标区域,更好地避免出现梯度消失问题,捕捉特征更加充分,有效提高分割性能和网络的可靠性。
- Abstract:
- In the past two years, the outbreak of new coronary pneumonia in the world has brought extremely serious security risks to human beings. In order to improve the efficiency of doctors in diagnosis as much as possible, it is of great value to study the lesion segmentation method of new coronary pneumonia images. This paper uses the U-Net network model as the basic network to process the pneumonia images, before the first and second layer splicing, the improved channel attention is introduced to strengthen important information; at the same time, before the third and fourth layers are spliced, the spatial attention is introduced to extract the spatial information, and the detailed information of the target area is supplemented. Finally, the mixed loss function is used to accelerate the convergence speed of the network and avoid the sample imbalance. Compared with the basic network, the improved network model can segment both large target areasand small target areas, and avoid the problem of gradient disappearance, and effectively improve the segmentation performance and the reliability of the network.
参考文献/References:
[1] Rt A.Pandemic potential of 2019-nCoV[J].The Lancet Infectious Diseases,2020,20(3):280-280.
[2] 李士雪,单莹.新型冠状病毒肺炎研究进展述评[J].山东大学学报(医学版),2020,58(3):19-25.
[3] 宋瑶,刘俊.改进U-Net的新冠肺炎图像分割方法[J].计算机工程与应用,2021,57(19):243-251.
[4] Razavian A S,Azizpour H,Sullivan J,et al.CNN Features off-the-shelf: an Astounding Baseline for Recognition[C].2014 IEEE conference on computer vision and pattern recognition workshops.IEEE,2014:512-519.
[5] Long J,Shelhamer E,Darrell T.Fully Convolutional Networks for Semantic Segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,39(4):640-651.
[6] Ronneberger O,Fischer P,Brox T.U-net: Convolutional networks for biomedical image segmentation[C].International Conference on Medical image computing and computer-assisted intervention.Springer,Cham,2015:234-241.
[7] Rajamani K T,Siebert H,Heinrich M P.Dynamic deformable attention network(DDANet)for COVID-19 lesions semantic segmentation[J].Journal of Biomedical Informatics,2021:103816.
[8] Zhao X,Zhang P,Song F,et al.D2a u-net: Automatic segmentation of covid-19 lesions from ct slices with dilated convolution and dual attention mechanism[J].arXiv preprint arXiv,2021(2):05210.
[9] 席佳妮.智能视觉物联网中基于车辆多属性检索的应用研究[D].呼和浩特:内蒙古大学,2021.
[10] Woo S,Park J,Lee J Y,et al.Cbam:Convolutional block attention module[C].Proceedings of the European conference on computer vision(ECCV),2018:3-19.
[11] Morozov S P,Andreychenko A E,Pavlov N A,et al.Mosmeddata:Chest ct scans with covid-19 related findings dataset[J].arXiv preprint arXiv,2020(5):06465.
[12] 付顺兵,王朝斌,罗建,等.基于改进U-Net模型的脑肿瘤MR图像分割[J].西华师范大学学报(自然科学版),2021,42(2):202-208.
[13] 莫春梅,周金治,李雪,等.基于改进U-Net的肝脏分割方法[J].中国医学物理学杂志,2021,38(5):571-577.
[14] 殷晓航,王永才,李德英.基于U-Net结构改进的医学影像分割技术综述[J].软件学报,2021,32(2):519-550.
[15] 李婧.基于深度学习的极化SAR图像去噪与分类研究[D].阜新:辽宁工程技术大学,2021.
[16] Xie E,Wang W,Yu Z,et al.SegFormer:Simple and efficient design for semantic segmentation with transformers[J].Advances in Neural Information Processing Systems,2021,34:12077-12090.
备注/Memo
收稿日期:2022-04-27
基金项目:四川省教育厅重点资助项目(14ZA0123)