[1]龚 安,赵 莉,姚鑫杰.基于改进的 U-Net 网络模型的气胸分割算法[J].计算机技术与发展,2021,31(10):173-178.[doi:10. 3969 / j. issn. 1673-629X. 2021. 10. 029]
　GONG An,ZHAO Li,YAO Xin-jie.Pneumothorax Segmentation Algorithm Based on Improved U-Net Network Model[J].,2021,31(10):173-178.[doi:10. 3969 / j. issn. 1673-629X. 2021. 10. 029]
点击复制

基于改进的 U-Net 网络模型的气胸分割算法()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
31
期数:
2021年10期
页码:
173-178
栏目:
应用前沿与综合
出版日期:
2021-10-10

文章信息/Info

Title:
Pneumothorax Segmentation Algorithm Based on Improved U-Net Network Model
文章编号:
1673-629X(2021)10-0173-06
作者:
龚 安; 赵 莉; 姚鑫杰
中国石油大学(华东) 计算机科学与技术学院,山东 青岛 266580
Author(s):
GONG An; ZHAO Li; YAO Xin-jie
School of Computer Science and Technology,China University of Petroleum ( East China) ,Qingdao 266580,China
关键词:
气胸; U-Net; 残差学习; 语义分割; 全连接条件随机场
Keywords:
pneumothorax; U-Net; residual learning; semantic segmentation; fully connected conditional random field
分类号:
TP391
DOI:
10. 3969 / j. issn. 1673-629X. 2021. 10. 029
摘要:
X 线胸片图像本身十分复杂,其组织对比度低,器官组织之间的边界形态不规则,再加上气胸病灶区域特征不明显, 诊断严重依赖放射科医生的经验,传统分割算法常常需要人工干预,不能实现全自动分割病灶区域。 针对以上问题,提出一种改进 U-Net 的网络模型算法,实现自动化分割气胸。该网络保持编解码体系结构,将 U-Net 编码器中的结构替换为 ResNet 结构,引入残差学习模块提高特征学习能力,解码器采用卷积和上采样将特征图恢复到原图像大小,编码器-解码器之间依旧采用 U-Net 结构的特征融合方式拼接。 训练阶段为加快网络收敛速度调用 ResNet50 预训练参数,对模型预测的分割结果利用全连接条件随机场做图像后处理。 实验结果表明, 该算法有效提升了深度学习在处理气胸分割任务上的分割精度, Dice 相似系数稳定在 0. 851,Jaccard 系数稳定在 0. 769。
Abstract:
The chest X-ray image itself is quite complex,with low tissue contrast and irregular boundary morphology between organs and tissues. In addition,the characteristics of pneumothorax lesion area are not significant,so the diagnosis depends heavily on the experience of radiologists. The traditional segmentation algorithm often needs manual intervention,and cannot fully automatically segment the lesion area. In view of the above problems,an improved U-Net network model algorithm is proposed to realize the function of automatic pneumothorax segmentation. The network maintains the coding and decoding architecture,replaces the structure in the U-Net encoder with the ResNet structure,and introduces a residual learning module to improve the feature learning ability. The decoder uses the deconvolution network to restore the feature map to the original image size. The encoder and decoder are still spliced using the feature fusion method ofU-Net structure. In the training stage, ResNet50 pre-training parameters are invoked to speed up the network convergence. The segmentation result predicted by the model is subjected to image post-processing using the fully connected conditional random field. The experiment shows that the proposed algorithm effectively improves the segmentation accuracy of deep learning in dealing with pneumothorax segmentation tasks. The Dice similarity coefficient is stable at 0. 851,and the Jaccard coefficient is stable at 0. 769.

相似文献/References:

[1]刘　鑫,赵慧杰,武苏雯.基于Ｕ－Ｎｅｔ的岩层图像裂缝检测算法应用[J].计算机技术与发展,2021,31(05):73.[doi:10. 3969 / j. issn. 1673-629X. 2021. 05. 013]
　,,et al.ＡｐｐｌｉｃａｔｉｏｎｏｆＵ－ＮｅｔＢａｓｅｄＲｏｃｋＦｏｒｍａｔｉｏｎＩｍａｇｅＣｒａｃｋＤｅｔｅｃｔｉｏｎＡｌｇｏｒｉｔｈｍ[J].,2021,31(10):73.[doi:10. 3969 / j. issn. 1673-629X. 2021. 05. 013]
[2]李博文,唐贵进,崔子冠.基于Ｒｅｔｉｎｅｘ模型的弱光照图像增强算法[J].计算机技术与发展,2021,31(05):79.[doi:10. 3969 / j. issn. 1673-629X. 2021. 05. 014]
　,ＡＷｅａｋｌｙＩｌｌｕｍｉｎａｔｅｄＩｍａｇｅＥｎｈａｎｃｅｍｅｎｔＡｌｇｏｒｉｔｈｍ ＢａｓｅｄｏｎＲｅｔｉｎｅｘＭｏｄｅｌ[J].,2021,31(10):79.[doi:10. 3969 / j. issn. 1673-629X. 2021. 05. 014]
[3]史健婷,崔闫靖,常 亮.基于优化 U-Net 网络的乳腺肿瘤区域分割方法[J].计算机技术与发展,2021,31(08):156.[doi:10. 3969 / j. issn. 1673-629X. 2021. 08. 027]
　SHI Jian-ting,CUI Yan-jing,CHANG Liang.A Segmentation Method of Breast Tumor Region Based onOptimized U-Net[J].,2021,31(10):156.[doi:10. 3969 / j. issn. 1673-629X. 2021. 08. 027]
[4]喻鲁立,陈 黎.基于 U-Net 的多尺度视网膜血管分割方法[J].计算机技术与发展,2022,32(04):140.[doi:10. 3969 / j. issn. 1673-629X. 2022. 04. 024]
　YU Lu-li,CHEN Li.Multi-scale Segmentation of Retinal Vessels Based on U-Net[J].,2022,32(10):140.[doi:10. 3969 / j. issn. 1673-629X. 2022. 04. 024]
[5]黄 莉,何美玲*.基于 U-Net 改进模型的多模态脑肿瘤分割方法[J].计算机技术与发展,2022,32(11):58.[doi:10. 3969 / j. issn. 1673-629X. 2022. 11. 009]
　HUANG Li,HE Mei-ling*.Multi-model Brain Tumor Segmentation Method Based on Improved U-Net Model[J].,2022,32(10):58.[doi:10. 3969 / j. issn. 1673-629X. 2022. 11. 009]
[6]季倓正,邵允学,吕摇 刚.基于深度学习的钢坯入炉前定位技术[J].计算机技术与发展,2022,32(S2):36.[doi:10. 3969 / j. issn. 1673-629X. 2022. S2. 006]
　JI Tan-zheng,SHAO Yun-xue,LYU Gang.Positioning Technology of Billet before Furnace Based on Deep Learning[J].,2022,32(10):36.[doi:10. 3969 / j. issn. 1673-629X. 2022. S2. 006]
[7]罗 杨,万黎明,李 理,等.基于改进 U-Net 网络的半监督裂缝分割方法[J].计算机技术与发展,2022,32(12):179.[doi:10. 3969 / j. issn. 1673-629X. 2022. 12. 027]
　LUO Yang,WAN Li-ming,LI Li,et al.A Semi-supervised Crack Segmentation Method Based on Improved U-Net Network[J].,2022,32(10):179.[doi:10. 3969 / j. issn. 1673-629X. 2022. 12. 027]
[8]李秀华,朱水成.一种基于改进 U-Net 的肝脏肿瘤分割方[J].计算机技术与发展,2023,33(02):71.[doi:10. 3969 / j. issn. 1673-629X. 2023. 02. 011]
　LI Xiu-hua,ZHU Shui-cheng.A Liver Tumor Segmentation Method Based on Improved U-Net[J].,2023,33(10):71.[doi:10. 3969 / j. issn. 1673-629X. 2023. 02. 011]
[9]王宇昕,付晓薇,赵思宇,等.基于多层融合注意力的乳腺肿瘤图像分割方法[J].计算机技术与发展,2023,33(07):139.[doi:10. 3969 / j. issn. 1673-629X. 2023. 07. 021]
　WANG Yu-xin,FU Xiao-wei,ZHAO Si-yu,et al.Breast Tumor Images Segmentation Method Based on Multi-layer Fusion Attention[J].,2023,33(10):139.[doi:10. 3969 / j. issn. 1673-629X. 2023. 07. 021]
[10]谢宏彪,刘志勤,王庆凤,等.基于半监督的 3D 肝脏 CT 自动分割方法研究[J].计算机技术与发展,2023,33(09):149.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 022]
　XIE Hong-biao,LIU Zhi-qin,WANG Qing-feng,et al.Automated Segmentation of 3D Liver CT Images Using Semi-supervision[J].,2023,33(10):149.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 022]

更新日期/Last Update: 2021-10-10