基于级联Dense-UNet和图割的肝脏肿瘤自动分割

杨振; 邸拴虎; 赵于前; 廖苗; 曾业战

doi:10.11999/JEIT210247

基于级联Dense-UNet和图割的肝脏肿瘤自动分割

doi: 10.11999/JEIT210247

杨振^{1, 2},
邸拴虎¹,
赵于前^{1, 3, ,},
廖苗^{1, 4},
曾业战⁵

1.
中南大学自动化学院长沙 410083
2.
中南大学湘雅医院肿瘤科长沙 410083
3.
湖南湘江人工智能学院长沙 410083
4.
湖南科技大学计算机科学与工程学院湘潭 411100
5.
湖南工业大学电气与信息工程学院株洲 412007

基金项目: 国家自然科学基金(62076256, 61772555, 61702179)，湖南省教育厅资助科研项目(20B239, 18C0497)，湖南省研究生科研创新项目(CX20200129)，湖南省自然科学基金(2021JJ30275)

详细信息

作者简介:
杨振：男，1974年生，副教授，硕士生导师，研究方向为医学图像处理、图像引导放疗、放疗智能化、精准放疗剂量学

邸拴虎：男，1987年生，博士生，研究方向为医学图像处理、机器学习、放疗智能化

赵于前：男，1973年生，博士，教授，博士生导师，研究方向为医学图像处理、模式识别、视频处理等

廖苗：女，1988年生，博士，副教授，硕士生导师，研究方向为医学图像处理、图像分割、模式识别

曾业战：男，1980年生，博士，讲师，硕士生导师，研究方向为医学图像处理、人工智能、模式识别

通讯作者:
赵于前　 zyq@csu.edu.cn

中图分类号: TN911.73; TP391.41
计量
- 文章访问数: 1887
- HTML全文浏览量: 2567
- PDF下载量: 275
- 被引次数: 0
出版历程
- 收稿日期: 2021-03-26
- 修回日期: 2021-08-24
- 网络出版日期: 2021-09-14
- 刊出日期: 2022-05-25

Automatic Liver Tumor Segmentation Based on Cascaded Dense-UNet and Graph Cuts

YANG Zhen^{1, 2},
DI Shuanhu¹,
ZHAO Yuqian^{1, 3
, ,},
LIAO Miao^{1, 4},
ZENG Yezhan⁵

1.
School of Automation, Central South University, Changsha 410083, China
2.
Department of Oncology, Xiangya Hospital, Central South University, Changsha 410083, China
3.
Hunan Xiangjiang Artificial Intelligence Academy, Changsha, 410083, China
4.
School of Computer Science and Engineering, Hunan University of Science and Technology, Xiangtan 411100, China
5.
School of Electrical and Information Engineering, Hunan University of Technology, Zhuzhou 412007, China

Funds: The National Natural Science Foundation of China (62076256, 61772555, 61702179), The Scientific Research Fund of Hunan Provincial Education Department (20B239,18C0497), The Postgraduate Scientific Research Innovation Project of Hunan Province (CX20200129), Hunan Provincial Natural Science Foundation of China (2021JJ30275)

摘要

摘要: 腹部CT图像肝脏肿瘤分割是进行肝脏疾病诊断、手术规划和放射治疗的重要前提。针对肝脏肿瘤灰度异质、纹理丰富、边界模糊等因素引起的分割困难，该文提出基于级联Dense-Unet和图割的自动精确鲁棒分割方法。首先运用级联的Dense-UNet获取肝脏肿瘤初始分割结果及感兴趣区域，然后利用图像像素级和区域级特征，分别构建可有效区分肿瘤与非肿瘤的灰度模型和概率模型，并将其融入图割能量函数，进一步精确分割感兴趣区域中的肿瘤组织。最后分别采用LiTS和3Dircadb公共数据库作为训练集与测试集进行实验，并与现有多种自动分割方法进行了比较。结果表明，提出方法可有效分割CT图像中灰度、形状、大小、位置各异的肝脏肿瘤，能提取更精确的肿瘤边界，尤其对于对比度低、边界模糊的肿瘤具有明显优势。
- CT图像 /
- 肿瘤分割 /
- Dense-UNet /
- 图割
Abstract: Liver tumor segmentation from abdominal CT image is an important prerequisite for liver disease diagnosis, surgical planning, and radiation therapy. However, the segmentation remains a challenging problem since the tumors in CT images generally have heterogeneous intensities, complicated textures, and ambiguous boundaries. To address this, an automatic, accurate, and robust segmentation method is proposed based on cascaded Dense-Unet and graph cuts. Firstly, the cascaded Dense-UNet is used to obtain liver tumor initial segmentation results as well as the tumor Regions Of Interest (ROIs). Then, an intensity model and a probability model are established respectively by utilizing pixel-wise and patch-wise features in order to distinguish between tumor and non-tumor, and these models are further integrated into the graph cuts energy function to segment the tumor from ROIs accurately. Finally, experiments are carried out on LiTS and 3Dircadb datasets, which are respectively used as training set and testing set, and this method is compared with many other existing automatic segmentation methods. Results demonstrate that the proposed method can segment liver tumors in CT images with different intensity, texture, shape and size more effectively and can extract the tumor boundaries more accurately than other methods, especially for the tumors with low contrasts and ambiguous boundaries.
- CT image /
- Tumor segmentation /
- Dense-UNet /
- Graph cuts

HTML全文

图 1 Dense-UNet网络结构

下载: 全尺寸图片幻灯片

图 2 基于级联Dense-UNet的分割框架

下载: 全尺寸图片幻灯片

图 3 从某一序列中随机挑选的CT切片肝脏肿瘤感兴趣区域提取结果示例

下载: 全尺寸图片幻灯片

图 4 原始肿瘤感兴趣区域示例

下载: 全尺寸图片幻灯片

图 5 灰度模型结果示例

下载: 全尺寸图片幻灯片

图 6 概率模型结果示例

下载: 全尺寸图片幻灯片

图 7 图的构建示意图

下载: 全尺寸图片幻灯片

图 8 图割分割结果示例

下载: 全尺寸图片幻灯片

图 9 肝脏分割网络训练过程

下载: 全尺寸图片幻灯片

图 10 肿瘤分割网络训练过程

下载: 全尺寸图片幻灯片

图 11 3Dircadb数据库部分切片分割结果示例

下载: 全尺寸图片幻灯片

图 12 采用不同标注进行训练得到的肿瘤分割结果

下载: 全尺寸图片幻灯片

图 13 不同方法分割结果比较

下载: 全尺寸图片幻灯片

表 1 本文方法(DU+GC)与DU的分割性能比较(均值±标准差)(%)

方法	Dice	Precision	Recall
DU	70±12	84±19	70±8
DU+GC	78±8	89±10	71±7

下载: 导出CSV

表 2 本文方法与其他现有方法的分割性能比较(均值±标准差)(%)

方法	发表时间	Dice	Precision	Recall
UNet^[12]	2015	68±11	82±12	69±18
FCNs^[16]	2017	59±20	78±16	51±23
DeepLabv3+^[17]	2018	67±18	79±19	64±21
UNet 3+^[18]	2020	69±20	67±20	70±15
TransUNet^[19]	2021	70±13	83±16	63±11
本文方法(DU+GC)		78±8	89±10	71±7

下载: 导出CSV

参考文献(19)

[1]	CAO Wei, CHEN Hongda, YU Yiwei, et al. Changing profiles of cancer burden worldwide and in China: A secondary analysis of the global cancer statistics 2020[J]. Chinese Medical Journal, 2021, 134(7): 783–791. doi: 10.1097/CM9.0000000000001474
[2]	LI Yang, ZHAO Yuqian, ZHANG Fan, et al. Liver segmentation from abdominal CT volumes based on level set and sparse shape composition[J]. Computer Methods and Programs in Biomedicine, 2020, 195: 105533. doi: 10.1016/j.cmpb.2020.105533
[3]	ALIRR O I. Deep learning and level set approach for liver and tumor segmentation from CT scans[J]. Journal of Applied Clinical Medical Physics, 2020, 21(10): 200–209. doi: 10.1002/acm2.13003
[4]	SIRIAPISITH T, KUSAKUNNIRAN W, and HADDAWY P. Pyramid graph cut: Integrating intensity and gradient information for grayscale medical image segmentation[J]. Computers in Biology and Medicine, 2020, 126: 103997. doi: 10.1016/j.compbiomed.2020.103997
[5]	KADOURY S, VORONTSOV E, and TANG An. Metastatic liver tumour segmentation from discriminant Grassmannian manifolds[J]. Physics in Medicine & Biology, 2015, 60(16): 6459–6478. doi: 10.1088/0031-9155/60/16/6459
[6]	FORUZAN A H and CHEN Yenwei. Improved segmentation of low-contrast lesions using sigmoid edge model[J]. International Journal of Computer Assisted Radiology and Surgery, 2016, 11(7): 1267–1283. doi: 10.1007/s11548-015-1323-x
[7]	LIU Liangliang, WU Fangxiang, WANG Yuping, et al. Multi-receptive-field CNN for semantic segmentation of medical images[J]. IEEE Journal of Biomedical and Health Informatics, 2020, 24(11): 3215–3225. doi: 10.1109/jbhi.2020.3016306
[8]	ISENSEE F, JAEGER P F, KOHL S A A, et al. nnU-Net: A self-configuring method for deep learning-based biomedical image segmentation[J]. Nature Methods, 2021, 18(2): 203–211. doi: 10.1038/s41592-020-01008-z
[9]	LI Xiaomeng, CHEN Hao, QI Xiaojuan, et al. H-DenseUNet: Hybrid densely connected UNet for liver and tumor segmentation from CT volumes[J]. IEEE Transactions on Medical Imaging, 2018, 37(12): 2663–2674. doi: 10.1109/tmi.2018.2845918
[10]	JIANG Huiyan, SHI Tianyu, BAI Zhiqi, et al. AHCNet: An application of attention mechanism and hybrid connection for liver tumor segmentation in CT volumes[J]. IEEE Access, 2019, 7: 24898–24909. doi: 10.1109/access.2019.2899608
[11]	LI Guoqing, ZHANG Meng, LI Jiaojie, et al. Efficient densely connected convolutional neural networks[J]. Pattern Recognition, 2021, 109: 107610. doi: 10.1016/j.patcog.2020.107610
[12]	RONNEBERGER O, FISCHER P, and BROX T. U-Net: Convolutional networks for biomedical image segmentation[C]. The 18th International Conference on Medical Image Computing and Computer-Assisted Intervention, Munich, Germany, 2015: 234–241. doi: 10.1007/978-3-319-24574-4_28.
[13]	LIAO Miao, ZHAO Yuqian, LIU Xiyao, et al. Automatic liver segmentation from abdominal CT volumes using graph cuts and border marching[J]. Computer Methods and Programs in Biomedicine, 2017, 143: 1–12. doi: 10.1016/j.cmpb.2017.02.015
[14]	BILIC P, CHRIST P F, VORONTSOV E, et al. The Liver Tumor Segmentation benchmark (LiTS)[EB/OL]. https://arxiv.org/abs/1901.04056, 2019.
[15]	SOLER L, HOSTETTLER A, AGNUS V, et al. 3D image reconstruction for comparison of algorithm database: A patient-specific anatomical and medical image database[EB/OL]. https://www.ircad.fr/research/3dircadb/, 2019.
[16]	SUN Changjian, GUO Shuxu, ZHANG Huimao, et al. Automatic segmentation of liver tumors from multiphase contrast-enhanced CT images based on FCNs[J]. Artificial Intelligence in Medicine, 2017, 83: 58–66. doi: 10.1016/j.artmed.2017.03.008
[17]	CHEN L C, ZHU Yukun, PAPANDREOU G, et al. Encoder-decoder with atrous separable convolution for semantic image segmentation[C]. The 15th European Conference on Computer Vision, Munich, Germany, 2018: 801–818. doi: 10.1007/978-3-030-01234-2_49.
[18]	HUANG Huimin, LIN Lanfen, TONG Ruofeng, et al. UNet 3+: A full-scale connected UNet for medical image segmentation[C]. 2020 IEEE International Conference on Acoustics, Speech and Signal Processing, Barcelona, Spain, 2020: 1055–1059. doi: 10.1109/ICASSP40776.2020.9053405.
[19]	CHEN Jieneng, LU Yongyi, YU Qihang, et al. TransUNet: Transformers make strong encoders for medical image segmentation[EB/OL].https://arxiv.org/abs/2102.04306, 2021.