搜索

x

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于心脏电影磁共振图像的一种新的右心室多图谱分割方法

苏新宇 王丽嘉 朱艳春

引用本文:
Citation:

基于心脏电影磁共振图像的一种新的右心室多图谱分割方法

苏新宇, 王丽嘉, 朱艳春

A new method of multi- atlas segmentation of right ventricle based on cardiac film magnetic resonance images

Su Xin-Yu, Wang Li-Jia, Zhu Yan-Chun
PDF
HTML
导出引用
  • 基于心脏电影磁共振图像的右心室(RV)分割, 对心脏疾病的诊疗及预后有着十分重要的意义. 右心室结构复杂, 传统图像分割方法始终未能达到较高的精度. 多图谱方法通过配准和融合来实现RV分割, 是近几年RV分割中的主要方法之一. 本文提出一种新的右心室多图谱分割方法, 能够实现RV的全自动准分割. 本文首先采用自适应仿射传播算法获取一系列图谱集, 并基于豪斯多夫距离和归一化互信息选择与目标图像最相似的图谱集; 然后, 依次采用多分辨率的仿射变换和Diffeomorphic demons算法将目标图像配准到最相似图谱集, 并将配准得到的形变场应用于标记图像获得粗分割结果; 最后, 本文采用COLLATE算法融合粗分割结果得到RV轮廓. 30例心脏电影磁共振数据被用于回顾性分析. 本文算法与放射专家手工分割的RV 相比, Dice指标和豪斯多夫距离的平均值分别为0.84, 11.46 mm; 舒张末期容积, 收缩末期容积, 射血分数的相关系数和偏差均值分别是0.94, 0.90, 0.86; 2.5113, –3.4783, 0.0341. 与卷积神经网络相比, 本文算法在收缩末期的分割精度更接近手动分割结果. 实验结果表明, 该方法从有效的图谱选择和基于多分辨率的Diffeomorphic demons算法的多级配准提高了右心室分割的精度, 有望应用于临床辅助诊断.
    Segmentation of right ventricle (RV) in a cine cardiac magnetic resonance (CMR) image is essential for the diagnosis and therapy of cardiac diseases. Traditional image segmentation methods fail to achieve high accuracy due to the complex structure of RV. Multi-atlas frame, which transforms the segmentation into registration and fusion, has become one of the main segmentation methods of RV in recent years. In this paper, we suggest a new multi-atlas frame for the automatical and accurate segmentation of RV. Firstly, an adaptive affinity propagation algorithm is used to obtain a series of atlases, in which the atlas set most similar to the target image based on hausdorff distance and normalized mutual information is selected. Then, the target image is registered onto the selected atlas by using multi-resolution strategy-based affine transform and Diffeomorphic demons algorithm to generate a deformation field, which is applied to the label image to obtain coarse segmentation results of RV. Finally, the Consensus Level, Labeler Accuracy and Truth Estimation (COLLATE) algorithm is used to fuse the coarse segmentation result to obtain the RV. The 30 cine CMR datasets are applied to the retrospective analysis. The comparison between RV value from the present algorithm and that from the manual segmentation shows that the average dice index and hausdorff distance are 0.84 and 11.46 mm, respectively, the correlation coefficients and deviation means of endo-diastolic volume, endo-systolic volume and ejection fraction are 0.94, 0.90, 0.86, and 2.5113, –3.4783, 0.0341, respectively. Compared with convolutional neural networks, the new multi-atlas frame has an endo-systolic volume close to the manual result. The results show that the suggested method improves the accuracy and robustness of segmentation of RV from the effective atlas selection and multi-resolution Diffeomorphic demons algorithm-based registration, and it promises to be applied to clinical diagnosis.
      通信作者: 王丽嘉, lijiawangmri@163.com
    • 基金项目: 深圳市基础研究项目(批准号: JCYJ20160429172357751)资助的课题
      Corresponding author: Wang Li-Jia, lijiawangmri@163.com
    • Funds: Project supported by the Shenzhen Provincial Research Foundation for Basic Research, China (Grant No. JCYJ20160429172357751)
    [1]

    World Health Organization, http://origin.who.int/mediacentrse/factsheets/fs317/en/ [2019−4−17]

    [2]

    胡盛寿, 高润霖, 刘力生, 朱曼璐, 王文, 王拥军, 吴兆苏, 李惠君, 顾东风, 杨跃进, 郑哲, 陈伟伟 2019 中国循环杂志 34 209Google Scholar

    Hu S S, Gao R L, Liu L S, Zhu M L, Wang W, Wang Y J, Wu Z S, Li H J, Gu D F, Yang Y J, Zheng Z, Chen W W 2019 Chin. Circ. J. 34 209Google Scholar

    [3]

    Haddad F, Hunt S A, Rosenthal D N, Murphy D J 2008 Circulation 117 1436Google Scholar

    [4]

    Kutty S, Shang Q, Joseph N, Kowallick J T, Schuster A, Steinmetz M, Danford D A, Beerbaum P, Sarikouch S 2017 Int. J. Cardiol. 248 136Google Scholar

    [5]

    Frey B J, Dueck D 2007 Science 315 972Google Scholar

    [6]

    张耀楠, 陈传慎, 康雁 2014 东北大学学报(自然科学版) 35 795Google Scholar

    Zhang Y N, Chen C S, Kang Y 2014 J. Northeast. Univ. 35 795Google Scholar

    [7]

    王开军, 张军英, 李丹, 张新娜, 郭涛 2007 自动化学报 12 1242

    Wang K J, Zhang J Y, Li D, Zhang X N, Guo T 2007 Acta Automatica Sin. 12 1242

    [8]

    Zuluaga M A, Cardoso M J, Modat M, Ourselin S 2013 International Conference, FIMH London, UK, June 20−22, 2013 p174

    [9]

    Lorenzovaldes M, Sanchezortiz G I, Mohiaddin R H, Rueckert D 2002 MICCAI Tokyo, Japan September 25−28, 2002 p642

    [10]

    Bai W, Shi W, Oregan D P, Tong T, Wang H, Jamilcopley S, Peters N S, Rueckert D 2013 IEEE Trans. Med. Imaging 32 1302Google Scholar

    [11]

    Ou Y, Sotiras A, Paragios N, Davatzikos C 2011 Med. Image Anal. 15 622Google Scholar

    [12]

    Shi W Z, Zhuang X H, Pizarro L, Bai W J, Wang H Y, Tung K P, Edwards P J, Rueckert D 2012 MICCAI Nice, France October 1−5, 2012 p659

    [13]

    Thirion J 1998 Med. Image Anal. 2 243Google Scholar

    [14]

    周先春, 汪美玲, 周林锋, 吴琴 2015 物理学报 64 024205Google Scholar

    Zhou X C, Wang M L, Zhou L F, Wu Q 2015 Acta Phys. Sin. 64 024205Google Scholar

    [15]

    Vercauteren T, Pennec X, Perchant A, Ayache N 2009 NeuroImage 45 S61Google Scholar

    [16]

    Asman A J, Landman B A 2011 IEEE Trans. Med. Imaging 30 1779Google Scholar

    [17]

    王丽嘉, 苏新宇, 李亚, 胡立伟, 聂生东 2018 波谱学杂志 35 407Google Scholar

    Wang L J, Su X Y, Li Y, Hu L W, Nie S D 2018 Chin. J. Magn. Reson. 35 407Google Scholar

    [18]

    Awate S P, Zhu P, Whitaker R T 2012 MICCAI Nice, France October 1−5, 2012 p 103

    [19]

    Kyhl K, Ahtarovski K A, Iversen K, Thomsen C, Vejlstrup N, Engstrom T, Madsen P 2013 Am. J. Physiol Heart Circ. Physiol. 305 H1004Google Scholar

    [20]

    Lei X L, Liu H, Han Y C, Cheng W, Sun J Y, Luo Y, Yang D, Dong Y, Chung Y C, Chen Y H 2017 J. Magn. Reson. Imaging 45 1684Google Scholar

  • 图 1  结合多图谱和Diffeomorphic demons算法的右心室分割流程图

    Fig. 1.  Flow diagram of right ventricular segmentation combined with multi-atlas and Diffeomorphic demons algorithm.

    图 2  图谱图像聚类流程图

    Fig. 2.  Atlas image clustering flow chart

    图 3  四个右心室配准的典型例子 (a) 目标图像; (b) 图谱图像; (b) 标记图像; (d) 配准结果

    Fig. 3.  Four examples of typical right ventricular registration: (a) Target image; (b) atlas image; (c) label image; (d) registration result.

    图 4  RV标记图像的融合过程

    Fig. 4.  RV label image fusion process

    图 5  ED从基底到顶端的RV分割结果

    Fig. 5.  RV contour at ED from base to apex.

    图 6  ES从基底到顶端的RV分割结果

    Fig. 6.  RV contour at ES from base to apex.

    图 7  Dice指标的箱形图

    Fig. 7.  Box diagram of the Dice index.

    图 8  本文算法与金标准结果分析 相关性分析 (a) 舒张末期容积; (b) 收缩末期容积; (c) 射血分数; Bland-Altman分析 (d) 舒张末期容积; (e) 收缩末期容积; (f) 射血分数

    Fig. 8.  Analysis of algorithm and gold standard results. Correlation analysis (a) EDV; (b) ESV; (c) EF; Bland-Altman analysis (d) EDV; (e) ESV; (f) EF.

    图 9  深度学习与金标准结果分析 相关性分析 (a) 舒张末期容积; (b) 收缩末期容积; (c) 射血分数; Bland-Altman分析 (d) 舒张末期容积; (e) 收缩末期容积; (f) 射血分数

    Fig. 9.  Analysis of deep learning and gold standard results. Correlation analysis (a) EDV; (b) ESV; (c) EF; Bland-Altman analysis (d) EDV; (e) ESV; (f) EF.

    表 1  ED, ES的平均Dice指数和豪斯多夫距离

    Table 1.  Average Dice index and Hausdorff distance of ED and ES.

    DiceHD/mm
    ED0.87(0.10)10.76(4.5)
    ES0.81(0.16)12.16(6.54)
    下载: 导出CSV
  • [1]

    World Health Organization, http://origin.who.int/mediacentrse/factsheets/fs317/en/ [2019−4−17]

    [2]

    胡盛寿, 高润霖, 刘力生, 朱曼璐, 王文, 王拥军, 吴兆苏, 李惠君, 顾东风, 杨跃进, 郑哲, 陈伟伟 2019 中国循环杂志 34 209Google Scholar

    Hu S S, Gao R L, Liu L S, Zhu M L, Wang W, Wang Y J, Wu Z S, Li H J, Gu D F, Yang Y J, Zheng Z, Chen W W 2019 Chin. Circ. J. 34 209Google Scholar

    [3]

    Haddad F, Hunt S A, Rosenthal D N, Murphy D J 2008 Circulation 117 1436Google Scholar

    [4]

    Kutty S, Shang Q, Joseph N, Kowallick J T, Schuster A, Steinmetz M, Danford D A, Beerbaum P, Sarikouch S 2017 Int. J. Cardiol. 248 136Google Scholar

    [5]

    Frey B J, Dueck D 2007 Science 315 972Google Scholar

    [6]

    张耀楠, 陈传慎, 康雁 2014 东北大学学报(自然科学版) 35 795Google Scholar

    Zhang Y N, Chen C S, Kang Y 2014 J. Northeast. Univ. 35 795Google Scholar

    [7]

    王开军, 张军英, 李丹, 张新娜, 郭涛 2007 自动化学报 12 1242

    Wang K J, Zhang J Y, Li D, Zhang X N, Guo T 2007 Acta Automatica Sin. 12 1242

    [8]

    Zuluaga M A, Cardoso M J, Modat M, Ourselin S 2013 International Conference, FIMH London, UK, June 20−22, 2013 p174

    [9]

    Lorenzovaldes M, Sanchezortiz G I, Mohiaddin R H, Rueckert D 2002 MICCAI Tokyo, Japan September 25−28, 2002 p642

    [10]

    Bai W, Shi W, Oregan D P, Tong T, Wang H, Jamilcopley S, Peters N S, Rueckert D 2013 IEEE Trans. Med. Imaging 32 1302Google Scholar

    [11]

    Ou Y, Sotiras A, Paragios N, Davatzikos C 2011 Med. Image Anal. 15 622Google Scholar

    [12]

    Shi W Z, Zhuang X H, Pizarro L, Bai W J, Wang H Y, Tung K P, Edwards P J, Rueckert D 2012 MICCAI Nice, France October 1−5, 2012 p659

    [13]

    Thirion J 1998 Med. Image Anal. 2 243Google Scholar

    [14]

    周先春, 汪美玲, 周林锋, 吴琴 2015 物理学报 64 024205Google Scholar

    Zhou X C, Wang M L, Zhou L F, Wu Q 2015 Acta Phys. Sin. 64 024205Google Scholar

    [15]

    Vercauteren T, Pennec X, Perchant A, Ayache N 2009 NeuroImage 45 S61Google Scholar

    [16]

    Asman A J, Landman B A 2011 IEEE Trans. Med. Imaging 30 1779Google Scholar

    [17]

    王丽嘉, 苏新宇, 李亚, 胡立伟, 聂生东 2018 波谱学杂志 35 407Google Scholar

    Wang L J, Su X Y, Li Y, Hu L W, Nie S D 2018 Chin. J. Magn. Reson. 35 407Google Scholar

    [18]

    Awate S P, Zhu P, Whitaker R T 2012 MICCAI Nice, France October 1−5, 2012 p 103

    [19]

    Kyhl K, Ahtarovski K A, Iversen K, Thomsen C, Vejlstrup N, Engstrom T, Madsen P 2013 Am. J. Physiol Heart Circ. Physiol. 305 H1004Google Scholar

    [20]

    Lei X L, Liu H, Han Y C, Cheng W, Sun J Y, Luo Y, Yang D, Dong Y, Chung Y C, Chen Y H 2017 J. Magn. Reson. Imaging 45 1684Google Scholar

  • [1] 赵地, 赵莉芝, 甘永进, 覃斌毅. 基于支撑先验与深度图像先验的无预训练磁共振图像重建方法. 物理学报, 2022, 71(5): 058701. doi: 10.7498/aps.71.20211761
    [2] 赵地, 赵莉芝, 甘永进, 覃斌毅. 基于支撑先验与深度图像先验的无预训练磁共振图像重建方法. 物理学报, 2021, (): . doi: 10.7498/aps.70.20211761
    [3] 王甜甜, 王慧, 朱艳春, 王丽嘉. 基于位移流U-Net和变分自动编码器的心脏电影磁共振图像左心肌运动追踪. 物理学报, 2021, 70(22): 228701. doi: 10.7498/aps.70.20210885
    [4] 张夫一, 葛曼玲, 郭志彤, 谢冲, 杨泽坤, 宋子博. 静息态功能磁共振成像评估健康老年人认知行为 的多尺度熵模型研究. 物理学报, 2020, (): 008700. doi: 10.7498/aps.69.20200051
    [5] 张夫一, 葛曼玲, 郭志彤, 谢冲, 杨泽坤, 宋子博. 静息态功能磁共振成像评估健康老年人认知行为的多尺度熵模型研究. 物理学报, 2020, 69(10): 108703. doi: 10.7498/aps.69.20200050
    [6] 刘杰, 张建勋, 代煜. 基于多引导滤波的图像增强算法. 物理学报, 2018, 67(23): 238701. doi: 10.7498/aps.67.20181425
    [7] 王小艳, 汪芃, 唐国宁. 人类心室组织中波斑图引起的后除极化研究. 物理学报, 2017, 66(6): 068201. doi: 10.7498/aps.66.068201
    [8] 吴量, 陈方, 黄重阳, 丁国辉, 丁义明. 基于改进非线性拟合的核磁共振T2谱多指数反演. 物理学报, 2016, 65(10): 107601. doi: 10.7498/aps.65.107601
    [9] 马振鹤, 窦世丹, 马毓姝, 刘健, 赵玉倩, 刘江红, 吕江涛, 王毅. 基于光学相干层析成像的早期鸡胚心脏径向应变测量. 物理学报, 2016, 65(23): 235202. doi: 10.7498/aps.65.235202
    [10] 范虹, 韦文瑾, 朱艳春. 基于二维集合经验模式分解的距离正则化水平集磁共振图像分割. 物理学报, 2016, 65(16): 168701. doi: 10.7498/aps.65.168701
    [11] 周先春, 汪美玲, 周林锋, 吴琴. 基于Demons算法改进的图像去噪模型研究. 物理学报, 2015, 64(2): 024205. doi: 10.7498/aps.64.024205
    [12] 朱俊杰, 蒋式勤, 王伟远, 赵晨, 王永良, 李文生, 权薇薇. 多腔体心脏磁场模型的研究与应用. 物理学报, 2014, 63(5): 058703. doi: 10.7498/aps.63.058703
    [13] 范虹, 朱艳春, 王芳梅, 张旭梅. 多分辨率水平集算法的乳腺MR图像分割. 物理学报, 2014, 63(11): 118701. doi: 10.7498/aps.63.118701
    [14] 张品, 梁艳梅, 常胜江, 范海伦. 基于能量最小化的肾脏计算断层扫描图像分割方法. 物理学报, 2013, 62(20): 208701. doi: 10.7498/aps.62.208701
    [15] 宋长新, 马克, 秦川, 肖鹏. 结合稀疏编码和空间约束的红外图像聚类分割研究. 物理学报, 2013, 62(4): 040702. doi: 10.7498/aps.62.040702
    [16] 吴一全, 张金矿. 二维直方图θ划分最大Shannon熵图像阈值分割. 物理学报, 2010, 59(8): 5487-5495. doi: 10.7498/aps.59.5487
    [17] 唐英干, 邸秋艳, 赵立兴, 关新平, 刘福才. 基于二维最小Tsallis交叉熵的图像阈值分割方法. 物理学报, 2009, 58(1): 9-15. doi: 10.7498/aps.58.9
    [18] 潘克家, 陈 华, 谭永基. 基于差分进化算法的核磁共振T2谱多指数反演. 物理学报, 2008, 57(9): 5956-5961. doi: 10.7498/aps.57.5956
    [19] 梁艳梅, 翟宏琛, 常胜江, 张思远. 基于最大隶属度原则的彩色图像分割方法. 物理学报, 2003, 52(11): 2655-2659. doi: 10.7498/aps.52.2655
    [20] 刘爱琢, 裴奉奎. 自旋置换对称体系多脉冲及二维核磁共振实验的密度算符描述(Ⅱ)——多量子积算符方法. 物理学报, 1990, 39(8): 154-161. doi: 10.7498/aps.39.154-2
计量
  • 文章访问数:  10552
  • PDF下载量:  64
  • 被引次数: 0
出版历程
  • 收稿日期:  2019-04-21
  • 修回日期:  2019-07-02
  • 上网日期:  2019-10-01
  • 刊出日期:  2019-10-05

/

返回文章
返回