针对CT图像的肺实质分割中由边界粘连型肿瘤造成的肺边界缺陷修复问题,提出了一种基于三维曲面重建的修复方法.对肺实质边界曲率变化较大处的缺陷,二维图像上无法获得足够多的特征对肺实质边界进行修复.本文方法首先使用质心灰度法改进了三维区域生长算法,提取肺实质进行三维重建.再使用阈值法提取分布在缺陷周围的三维点云,对三维点云进行曲面重建即可得到完整的肺实质轮廓.实验结果表明:该方法与传统的凸包算法和滚球法相比,能够更加完整有效地修复边界粘连型肺实质边界的缺陷.
To solve the problem of restoring the lung boundary defects caused by the lung boundary adhesion tumor in CT image segmentation of the lung parenchyma, a method based on three-dimensional surface reconstruction is proposed. For defects with large curvature of the lung parenchymal boundary, the two-dimensional image cannot obtain enough features to restore the lung parenchymal boundary. In this paper, firstly, the centroid-gray algorithm is used to improve the three-dimensional region growth algorithm, and the lung parenchyma is extracted for three-dimensional reconstruction. Then the threshold method is used to extract the three-dimensional point cloud distributed around the defects, a complete outline of the lung parenchyma can be obtained by the surfase reconstruction of the three-dimensional point cloud. The experimental results show that compared with the traditional convex hull algorithm and the rolling ball method, this method is more complete and effective to restore the defects of lung parenchymal boundary with adhesion tumors.
[1] ALLEMANI C, MATSUDA T, CARLO V D, et al. Global surveillance of trends in cancer survival 2000-14(CONCO-RD-3):analysis of individual records for 37513025 patients diagnosed with one of 18 cancers from 322 population-based registries in 71 countries[J]. Lancet, 2018, 391(10125):1023-1075. DOI:10.1016/S0140-6736(17)33326-3.
[2] KUHNIGK J M, DICKEN V, BORNEMANN L, et al. Morphological segmentation and partial volume analysis for volumetry of solid pulmonary lesions in thoracic CT scans[J]. IEEE Trans Med Imaging, 2006, 25(4): 417-434. DOI:10.1109/tmi.2006.871547.
[3] JOHNSON P B, YOUNG L A, LAMICHHANE N, et al. Quantitative imaging: Correlating image features with the segmentation accuracy of PET based tumor contours in the lung[J]. Radiother Oncol, 2017, 123(2): 257-262. DOI:10.1016/j.radonc.2017.03.008.
[4] RITI Y F, NUGROHO H A, WIBIRAMA S, et al. Feature extraction for lesion margin characteristic classification fr-om CT Scan lungs image[C] //2016 1st International Conference on Information Technology, Information Systems and Electrical Engineering(ICITISEE), IEEE, 2017, 54-58. DOI:10.1109/ICITISEE.2016.7803047.
[5] 龚敬,王丽嘉,王远军,等.基于灰度积分投影与模糊C均值聚类的肺实质分割[J].中国生物医学工程学报,2015, 34(1): 109-113. DOI:10.3969/j.issn.0258-8021.2015.01.015.
[6] KHWAIRAKPAM A, HAZARIKA R A, KANDAR D. Image segmentation by fuzzy edge detection and region growing technique[M] //Lecture Notes in Electrical Engineering, Singapore: Springer Singapore, 2019: 51-64. DOI:10.1007/978-981-13-7091-5_5.
[7] WEI Y, SHEN G, LI J J. A fully automatic method for lung parenchyma segmentation and repairing[J]. J Digit Imaging, 2013, 26(3): 483-495. DOI:10.1007/s10278-012-9528-9.
[8] 代双凤,吕科,翟锐, 等.基于3D区域增长法和改进的凸包算法相结合的全肺分割方法[J].电子与信息学报,2016, 38(9): 2358-2364. DOI:10.11999/JEIT151365.
[9] KAZHDAN M, BOLITHO M, HOPPE H, et al. Possion surface reconstruction[C] //Proc of Symposium on Geometry Processing, 2006, 7:61-70. DOI:10.1145/1281991.1281997.
[10] 罗海洋.基于改进MC算法的医学图像三维重建方法的研究[D]. 哈尔滨:哈尔滨理工大学,2019.
[11] MANSON J, PETROVA G, SCHAEFER S. Streaming surface reconstruction using wavelets[J]. Comput Graph Forum, 2008, 27(5): 1411-1420. DOI:10.1111/j.1467-8659.2008.01281.x. (
