[1]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):71-76.[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(02):71-76.[doi:10. 3969 / j. issn. 1673-629X. 2023. 02. 011]
点击复制

一种基于改进 U-Net 的肝脏肿瘤分割方()
分享到：

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

卷:
33
期数:
2023年02期
页码:
71-76
栏目:
媒体计算
出版日期:
2023-02-10

文章信息/Info

Title:
A Liver Tumor Segmentation Method Based on Improved U-Net
文章编号:
1673-629X(2023)02-0071-06
作者:
李秀华; 朱水成
长春工业大学 计算机科学与工程学院,吉林 长春 130012
Author(s):
LI Xiu-hua; ZHU Shui-cheng
School of Computer Science and Engineering,Changchun University of Technology,Changchun 130012,China
关键词:
肝脏肿瘤分割; U-Net; 混合空洞卷积; 密集上采样卷积; 残差模块; 注意力机制
Keywords:
liver tumor segmentation; U-Net; hybrid dilated convolution; dense upsampling convolution; residual module; attention mech鄄anism
分类号:
TP391
DOI:
10. 3969 / j. issn. 1673-629X. 2023. 02. 011
摘要:
肝脏肿瘤分割是肝癌诊断与治疗不可或缺的重要环节。 针对传统的 U-Net 网络在形状、大小、位置复杂多变且边界模糊的肿瘤分割中信息丢失、分割精度低等问题,对其进行改进以提高肝脏肿瘤分割精度。 首先,在编码阶段使用混合空洞卷积替换原有卷积块,增大感受野、获取更多的上下文信息;在解码阶段采用密集上采样卷积,捕获和解码更详细的信息;引入残差模块,加速模型的训练并防止网络退化。 其次,在每个跳跃连接之间加入注意力机制,使模型重点关注感兴趣区域,抑制冗余特征;使用组归一化( GN) 代替常用的批量归一化( BN) ,减小 Batch Size 过小对网络准确性的影响,并结合 Focal Tversky 损失函数以改善类不平衡问题。 通过 LiTS2017 数据集的实验表明,相较于传统 U-Net,所提改进模型在肝脏和肿瘤分割中的 Dice 指标分别提升了 3. 56% 和 4. 21% ,召回率提升了 3. 71% 和 5. 35% 。
Abstract:
Segmentation of liver tumor is an indispensable link in diagnosis and treatment of hepatocellular carcinoma. Aiming at theproblems of information loss and low segmentation accuracy of traditional U-Net network in tumor segmentation with complex shape,size and location and fuzzy boundary,it is improved to increase the accuracy of liver tumor segmentation. Firstly,in the encoding stage,the hybrid dilated convolution is used to replace the original convolution block to increase the receptive field and obtain more context information. In the decoding stage,dense upsampling convolution is used to capture and decode more detailed information. The residualmodule is introduced to speed up the training of the model and prevent network degradation. Secondly,the attention mechanism is addedbetween each jump connection to make the model focus on the region of interest and suppress redundant features. Group Normalization( GN) is used instead of common Batch Normalization ( BN) to reduce the impact of too small Batch Size on network accuracy,andcombined with Focal Tversky loss function to improve the class imbalance problem. Experiments on the LiTS2017 data set show thatcompared with the traditional U-Net,the Dice index of the proposed improved model in liver and tumor segmentation has increased by3.56% and 4. 21% respectively,while the recall rate has increased by 3. 71% and 5. 35% .

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(02):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(02):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(02):156.[doi:10. 3969 / j. issn. 1673-629X. 2021. 08. 027]
[4]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.[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(02):173.[doi:10. 3969 / j. issn. 1673-629X. 2021. 10. 029]
[5]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(02):140.[doi:10. 3969 / j. issn. 1673-629X. 2022. 04. 024]
[6]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(02):58.[doi:10. 3969 / j. issn. 1673-629X. 2022. 11. 009]
[7]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(02):36.[doi:10. 3969 / j. issn. 1673-629X. 2022. S2. 006]
[8]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(02):179.[doi:10. 3969 / j. issn. 1673-629X. 2022. 12. 027]
[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(02):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(02):149.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 022]

更新日期/Last Update: 2023-02-10