Deep Distance Transform for Tubular Structure Segmentation in CT Scans

CVPR, pp. 3832-3841, 2019.

Cited by: 8|Views162
EI
Weibo:
We present Deep Distance Transform for accurate tubular structure segmentation, which combines intuitions from the classical distance transform for skeletonization and modern deep segmentation networks

Abstract:

Tubular structure segmentation in medical images, e.g., segmenting vessels in CT scans, serves as a vital step in the use of computers to aid in screening early stages of related diseases. But automatic tubular structure segmentation in CT scans is a challenging problem, due to issues such as poor contrast, noise and complicated backgro...More

Code:

Data:

0
Introduction
  • Tubular structures are ubiquitous throughout the human body, with notable examples including blood vessels, pancreatic duct and urinary tract.
  • FCN and its variants have become out-of-the-box models for tubular organ/tissue segmentation and achieve state-of-the-art results [24, 46]
  • These networks try to learn a class label per voxel, which inevitably ignores the geometric arrangement of the voxels in a tubular structure, and can not guarantee that the obtained segmentation has the right shape.
  • How can the authors use dilated duct as a cue to help find the PDAC tumor in an abnormal case even if it does NOT have any PDAC tumor prediction by directly applying deep networks?
Highlights
  • Tubular structures are ubiquitous throughout the human body, with notable examples including blood vessels, pancreatic duct and urinary tract
  • We investigate automatic tubular organ/tissue segmentation from CT scans, which is important for the characterization of various diseases [18]
  • To tackle the obstacles mentioned above, we propose to perform tubular structure segmentation by training a multitask deep network to predict a segmentation mask for a tubular structure, and a distance map, consisting of the distance transform value from each tubular structure voxel to the tubular structure surface, rather than a single skeleton/non-skeleton label
  • We present Deep Distance Transform (DDT) for accurate tubular structure segmentation, which combines intuitions from the classical distance transform for skeletonization and modern deep segmentation networks
  • We evaluated our approach on six datasets including four tubular structures: pancreatic duct, aorta, veins and vessels
  • Experiment shows the superiority of the proposed Deep Distance Transform for tubular structure segmentation and clinical application
Methods
  • SegBaseline [46] Multi-phase HPN [46] DDT (Ours) V A+V V.
  • For 3D-UNet, the DDT even outperforms the multi-phase method by more than 13% in terms of DSC.
  • The authors conduct ablation experiments on the PDAC segmentation dataset, using ResDSN as the backbone.
  • These variants of the methods are considered:
Results
  • Results and Discussions

    To evaluate the performance of the proposed DDT framework, the authors compare it with a per-.
  • To evaluate the performance of the proposed DDT framework, the authors compare it with a per-
Conclusion
  • The authors present Deep Distance Transform (DDT) for accurate tubular structure segmentation, which combines intuitions from the classical distance transform for skeletonization and modern deep segmentation networks.
  • The authors evaluated the approach on six datasets including four tubular structures: pancreatic duct, aorta, veins and vessels.
  • Experiment shows the superiority of the proposed DDT for tubular structure segmentation and clinical application
Summary
  • Introduction:

    Tubular structures are ubiquitous throughout the human body, with notable examples including blood vessels, pancreatic duct and urinary tract.
  • FCN and its variants have become out-of-the-box models for tubular organ/tissue segmentation and achieve state-of-the-art results [24, 46]
  • These networks try to learn a class label per voxel, which inevitably ignores the geometric arrangement of the voxels in a tubular structure, and can not guarantee that the obtained segmentation has the right shape.
  • How can the authors use dilated duct as a cue to help find the PDAC tumor in an abnormal case even if it does NOT have any PDAC tumor prediction by directly applying deep networks?
  • Methods:

    SegBaseline [46] Multi-phase HPN [46] DDT (Ours) V A+V V.
  • For 3D-UNet, the DDT even outperforms the multi-phase method by more than 13% in terms of DSC.
  • The authors conduct ablation experiments on the PDAC segmentation dataset, using ResDSN as the backbone.
  • These variants of the methods are considered:
  • Results:

    Results and Discussions

    To evaluate the performance of the proposed DDT framework, the authors compare it with a per-.
  • To evaluate the performance of the proposed DDT framework, the authors compare it with a per-
  • Conclusion:

    The authors present Deep Distance Transform (DDT) for accurate tubular structure segmentation, which combines intuitions from the classical distance transform for skeletonization and modern deep segmentation networks.
  • The authors evaluated the approach on six datasets including four tubular structures: pancreatic duct, aorta, veins and vessels.
  • Experiment shows the superiority of the proposed DDT for tubular structure segmentation and clinical application
Tables
  • Table1: Performance comparison (DSC, %) on pancreatic duct segmentation (mean ± standard deviation of all cases). SegBaseline stands for per-voxel classification. Multi-phase HPN is a hyper-paring network combining CT scans from both venous (V) and arterial (A) phases. Noted that only CT scans in venous phase are used for SegBaseline and DDT. Bold denotes the best results
  • Table2: Ablation study of pancreatic duct segmentation using ResDSN as backbone network. GAR indicates the proposed geometry-aware refinement
  • Table3: Performance comparison (in average DSC, % and mean surface distance in mm) on three tubular structure datasets by using different backbones. “↑” and “↓” indicate the larger and the smaller the better, respectively. Bold denotes the best results for each tubular structure per measurement
  • Table4: Comparison to competing submissions of MSD challenge: http://medicaldecathlon.com
  • Table5: Normal vs. abnormal classification results. Zhu et al [<a class="ref-link" id="c49" href="#r49">49</a>] + ours denotes applying our method to find the missing tumor of Zhu et al.. “↑” and “↓” indicate the larger and the smaller the better, respectively
Download tables as Excel
Related work
  • 2.1. Tubular Structure Segmentation

    2.1.1 Geometry-based Methods

    Various methods have been proposed to improve the performance of tubular structure segmentation by considering the geometric characteristics, and a non-exhaustive overview is given here. (1) Contour-based methods extracted the segmentation mask of a tubular structure by means of approximating its shape in the cross-sectional domain [1, 10]. (2) Minimal path approaches conducted tubular structure tracking and were usually interactive. They captured the global minimum curve (energy weighted by the image potential) between two points given by the user [9]. (3) Modelbased tracking methods required to refine a tubular structure model, which most of the time adopted a 3D cylinder with elliptical or circular section. At each tracking step, they calculated the new model position by finding the best model match among all possible new model positions [8]. (4) Centerline based methods found the centerline and estimated the radius of linear structures. For example, multiscale centerline detection method proposed in [34] adopted the idea of distance transform, and reformulated centerline detection and radius estimation in terms of a regression problem in 2D. Our work fully leverages the geometric information of a tubular structure, proposing a distance transform algorithm to implicitly learn the skeleton and cross-sectional radius, and the final segmentation mask is reconstructed by adopt-
Funding
  • Proposes a geometryaware tubular structure segmentation method, Deep Distance Transform , which combines intuitions from the classical distance transform for skeletonization and modern deep segmentation networks
  • Investigates automatic tubular organ/tissue segmentation from CT scans, which is important for the characterization of various diseases
  • Proposes to perform tubular structure segmentation by training a multitask deep network to predict a segmentation mask for a tubular structure, and a distance map, consisting of the distance transform value from each tubular structure voxel to the tubular structure surface, rather than a single skeleton/non-skeleton label
  • Proposes a distance loss term used for network training, which indicates a penalty when predicted distance transform value is far away from its ground-truth
  • Aims at designing an integrated framework which mines traditional distance transform and model deep networks for such cylinder-like shape structure, which is not studied in prior research
Reference
  • Luis Alvarez, Agustın Trujillo-Pino, Carmelo Cuenca, Esther Gonzalez, Julio Escların, Luis Gomez, Luis Mazorra, Miguel Aleman-Flores, Pablo G. Tahoces, and Jose M. Carreira-Villamor. Tracking the aortic lumen geometry by optimizing the 3d orientation of its cross-sections. In Proc. MICCAI, 2017.
    Google ScholarLocate open access versionFindings
  • Roberto Annunziata, Ahmad Kheirkhah, Pedram Hamrah, and Emanuele Trucco. Scale and curvature invariant ridge detector for tortuous and fragmented structures. In Proc. MICCAI, pages 588–595, 2015.
    Google ScholarLocate open access versionFindings
  • Luca Antiga, Bogdan Ene-Iordache, and Andrea Remuzzi. Computational geometry for patient-specific reconstruction and meshing of blood vessels from angiography. IEEE Trans. Med. Imaging, 22(5):674–684, 2003.
    Google ScholarLocate open access versionFindings
  • Carlo Arcelli and Gabriella Sanniti di Baja. Ridge points in euclidean distance maps. Pattern Recognition Letters, 13(4):237–243, 1992.
    Google ScholarLocate open access versionFindings
  • Oscar Kin-Chung Au, Chiew-Lan Tai, Hung-Kuo Chu, Daniel Cohen-Or, and Tong-Yee Lee. Skeleton extraction by mesh contraction. ACM Trans. Graph., 27(3):44:1–44:10, 2008.
    Google ScholarLocate open access versionFindings
  • Andreas Baak, Meinard Muller, Gaurav Bharaj, Hans-Peter Seidel, and Christian Theobalt. A data-driven approach for real-time full body pose reconstruction from a depth camera. In Proc. ICCV, 2011.
    Google ScholarLocate open access versionFindings
  • Christian Bauer and Horst Bischof. Extracting curve skeletons from gray value images for virtual endoscopy. In International Workshop on Medical Imaging and Virtual Reality, pages 393–402, 2008.
    Google ScholarLocate open access versionFindings
  • Christian Bauer, Thomas Pock, Erich Sorantin, Horst Bischof, and Reinhard Beichel. Segmentation of interwoven 3d tubular tree structures utilizing shape priors and graph cuts. Medical Image Analysis, 14(2):172–184, 2010.
    Google ScholarLocate open access versionFindings
  • Fethallah Benmansour and Laurent D. Cohen. Tubular structure segmentation based on minimal path method and anisotropic enhancement. International Journal of Computer Vision, 92(2):192–210, 2011.
    Google ScholarLocate open access versionFindings
  • Vicent Caselles, Ron Kimmel, and Guillermo Sapiro. Geodesic active contours. International Journal of Computer Vision, 22(1):61–79, 1997.
    Google ScholarLocate open access versionFindings
  • Linda C. Chu., Michael G. Goggins., and Elliot K. Fishman. Diagnosis and detection of pancreatic cancer. The Cancer Journal, 23(6):333–342, 2017.
    Google ScholarLocate open access versionFindings
  • Ozgun Cicek, Ahmed Abdulkadir, Soeren S. Lienkamp, Thomas Brox, and Olaf Ronneberger. 3d u-net: Learning dense volumetric segmentation from sparse annotation. In Proc. MICCAI, 2016.
    Google ScholarLocate open access versionFindings
  • Qi Dou, Cheng Ouyang, Cheng Chen, Hao Chen, and PhengAnn Heng. Unsupervised cross-modality domain adaptation of convnets for biomedical image segmentations with adversarial loss. In Proc. IJCAI, 2018.
    Google ScholarLocate open access versionFindings
  • Mark D Edge, Maarouf Hoteit, Amil P Patel, Xiaoping Wang, Deborah A Baumgarten, and Qiang Cai. Clinical significance of main pancreatic duct dilation on computed tomography: Single and double duct dilation. World J Gastroenterol., 13(11):1701–1705, 2007.
    Google ScholarLocate open access versionFindings
  • Maayan Frid-Adar, Rula Amer, and Hayit Greenspan. Endotracheal tube detection and segmentation in chest radiographs using synthetic data. In Proc. MICCAI, pages 784– 792, 2019.
    Google ScholarLocate open access versionFindings
  • Huazhu Fu, Yanwu Xu, Stephen Lin, Damon Wing Kee Wong, and Jiang Liu. Deepvessel: Retinal vessel segmentation via deep learning and conditional random field. In Proc. MICCAI, pages 132–139, 2016.
    Google ScholarLocate open access versionFindings
  • Yaorong Ge and J. Michael Fitzpatrick. On the generation of skeletons from discrete euclidean distance maps. IEEE Trans. Pattern Anal. Mach. Intell., 18(11):1055–1066, 1996.
    Google ScholarLocate open access versionFindings
  • Florent Grelard, Fabien Baldacci, Anne Vialard, and JeanPhilippe Domenger. New methods for the geometrical analysis of tubular organs. Medical Image Analysis, 42:89–101, 2017.
    Google ScholarLocate open access versionFindings
  • Fabian Isensee, Jens Petersen, Simon A. A. Kohl, Paul F. Jager, and Klaus H. Maier-Hein. nnu-net: Breaking the spell on successful medical image segmentation. CoRR, abs/1904.08128, 2019.
    Findings
  • Wei Ke, Jie Chen, Jianbin Jiao, Guoying Zhao, and Qixiang Ye. SRN: side-output residual network for object symmetry detection in the wild. In Proc. CVPR, 2017.
    Google ScholarLocate open access versionFindings
  • Tom Sie Ho Lee, Sanja Fidler, and Sven J. Dickinson. Detecting curved symmetric parts using a deformable disc model. In Proc. ICCV, 2013.
    Google ScholarLocate open access versionFindings
  • Alex Levinshtein, Cristian Sminchisescu, and Sven J. Dickinson. Multiscale symmetric part detection and grouping. International Journal of Computer Vision, 104(2):117–134, 2013.
    Google ScholarLocate open access versionFindings
  • Jonathan Long, Evan Shelhamer, and Trevor Darrell. Fully Convolutional Networks for Semantic Segmentation. Proc. CVPR, 2015.
    Google ScholarLocate open access versionFindings
  • Jameson Merkow, Alison Marsden, David J. Kriegman, and Zhuowen Tu. Dense Volume-to-Volume Vascular Boundary Detection. arXiv preprint arXiv:1605.08401, 2016.
    Findings
  • Delphine Nain, Anthony J. Yezzi, and Greg Turk. Vessel segmentation using a shape driven flow. In Proc. MICCAI, 2004.
    Google ScholarLocate open access versionFindings
  • Mathias Perslev, Erik Bjørnager Dam, Akshay Pai, and Christian Igel. One network to segment them all: A general, lightweight system for accurate 3d medical image segmentation. In MICCAI, 2019.
    Google ScholarLocate open access versionFindings
  • Wayne Rosamond, Katherine Flegal, Karen Furie, Alan Go, Kurt Greenlund, Nancy Haase, Susan M. Hailpern, Michael Ho, Virginia Howard, Brett Kissela, Steven Kittner, Donald Lloyd-Jones, Mary McDermott, James Meigs, Claudia Moy, Graham Nichol, Christopher ODonnell, Veronique Roger, Paul Sorlie, Julia Steinberger, Thomas Thom, Matt Wilson, and Yuling Hong. Heart disease and stroke statistics-2008 update. Circulation, 117(4), 2008.
    Google ScholarLocate open access versionFindings
  • Azriel Rosenfeld and John L. Pfaltz. Distance functions on digital pictures. Pattern Recognition, 1:33–61, 1968.
    Google ScholarLocate open access versionFindings
  • Holger R. Roth, Hirohisa Oda, Yuichiro Hayashi, Masahiro Oda, Natsuki Shimizu, Michitaka Fujiwara, Kazunari Misawa, and Kensaku Mori. Hierarchical 3d fully convolutional networks for multi-organ segmentation. CoRR, abs/1704.06382, 2017.
    Findings
  • Rasmus Rothe, Radu Timofte, and Luc Van Gool. Deep expectation of real and apparent age from a single image without facial landmarks. International Journal of Computer Vision, 126(2-4):144–157, 2018.
    Google ScholarLocate open access versionFindings
  • Wei Shen, Xiang Bai, Zihao Hu, and Zhijiang Zhang. Multiple instance subspace learning via partial random projection tree for local reflection symmetry in natural images. Pattern Recognition, 52:306–316, 2016.
    Google ScholarLocate open access versionFindings
  • Wei Shen, Kai Zhao, Yuan Jiang, Yan Wang, Xiang Bai, and Alan L. Yuille. Deepskeleton: Learning multi-task scale-associated deep side outputs for object skeleton extraction in natural images. IEEE Trans. Image Processing, 26(11):5298–5311, 2017.
    Google ScholarLocate open access versionFindings
  • Amber L. Simpson, Michela Antonelli, Spyridon Bakas, Michel Bilello, Keyvan Farahani, Bram van Ginneken, Annette Kopp-Schneider, Bennett A. Landman, Geert J. S. Litjens, Bjoern H. Menze, Olaf Ronneberger, Ronald M. Summers, Patrick Bilic, Patrick Ferdinand Christ, Richard K. G. Do, Marc Gollub, Jennifer Golia-Pernicka, Stephan Heckers, William R. Jarnagin, Maureen McHugo, Sandy Napel, Eugene Vorontsov, Lena Maier-Hein, and M. Jorge Cardoso. A large annotated medical image dataset for the development and evaluation of segmentation algorithms. CoRR, abs/1902.09063, 2019.
    Findings
  • Amos Sironi, Vincent Lepetit, and Pascal Fua. Multiscale centerline detection by learning a scale-space distance transform. In CVPR, 2014.
    Google ScholarLocate open access versionFindings
  • Joao V. B. Soares, Jorge J. G. Leandro, Roberto M. Cesar, Herbert F. Jelinek, and Michael J. Cree. Retinal vessel segmentation using the 2-d gabor wavelet and supervised classification. IEEE Trans. Med. Imaging, 25(9):1214–1222, 2006.
    Google ScholarLocate open access versionFindings
  • Joes Staal, Michael D. Abramoff, Meindert Niemeijer, Max A. Viergever, and Bram van Ginneken. Ridge-based vessel segmentation in color images of the retina. IEEE Trans. Med. Imaging, 23(4):501–509, 2004.
    Google ScholarLocate open access versionFindings
  • Ahmed Taha, Pechin Lo, Junning Li, and Tao Zhao. Kid-net: Convolution networks for kidney vessels segmentation from ct-volumes. In Proc. MICCAI, 2018.
    Google ScholarLocate open access versionFindings
  • Stavros Tsogkas and Iasonas Kokkinos. Learning-based symmetry detection in natural images. In Proc. ECCV, pages 41–54, 2012.
    Google ScholarLocate open access versionFindings
  • Yukang Wang, Yongchao Xu, Stavros Tsogkas, Xiang Bai, Sven J. Dickinson, and Kaleem Siddiqi. Deepflux for skeletons in the wild. In Proc. CVPR, 2019.
    Google ScholarLocate open access versionFindings
  • Yan Wang, Yuyin Zhou, Wei Shen, Seyoun Park, Elliot K. Fishman, and Alan L. Yuille. Abdominal multi-organ segmentation with organ-attention networks and statistical fusion. Medical Image Analysis, 55:88–102, 2019.
    Google ScholarLocate open access versionFindings
  • Yan Wang, Yuyin Zhou, Peng Tang, Wei Shen, Elliot K. Fishman, and Alan L. Yuille. Training multi-organ segmentation networks with sample selection by relaxed upper confident bound. In Proc. MICCAI, 2018.
    Google ScholarLocate open access versionFindings
  • Onno Wink, Wiro J. Niessen, and Max A. Viergever. Fast delineation and visualization of vessels in 3d angiographic images. IEEE Trans. Med. Imaging, 19(4):337–346, 2000.
    Google ScholarLocate open access versionFindings
  • Yingda Xia, Fengze Liu, Dong Yang, Jinzheng Cai, Lequan Yu, Zhuotun Zhu, Daguang Xu, Alan L. Yuille, and Holger Roth. 3d semi-supervised learning with uncertainty-aware multi-view co-training. In WACV, 2019.
    Google ScholarLocate open access versionFindings
  • Peter J. Yim, Juan R. Cebral, Rakesh Mullick, and Peter L. Choyke. Vessel surface reconstruction with a tubular deformable model. IEEE Trans. Med. Imaging, 20(12):1411– 1421, 2001.
    Google ScholarLocate open access versionFindings
  • Kai Zhao, Wei Shen, Shanghua Gao, Dandan Li, and MingMing Cheng. Hi-fi: Hierarchical feature integration for skeleton detection. In Proc. IJCAI, 2018.
    Google ScholarLocate open access versionFindings
  • Yuyin Zhou, Yingwei Li, Zhishuai Zhang, Yan Wang, Angtian Wang, Elliot K Fishman, Alan L Yuille, and Seyoun Park. Hyper-pairing network for multi-phase pancreatic ductal adenocarcinoma segmentation. In Proc.MICCAI, 2019.
    Google ScholarLocate open access versionFindings
  • Yuyin Zhou, Lingxi Xie, Wei Shen, Yan Wang, Elliot K. Fishman, and Alan L. Yuille. A fixed-point model for pancreas segmentation in abdominal CT scans. In Proc. MICCAI, 2017.
    Google ScholarLocate open access versionFindings
  • Zhuotun Zhu, Yingda Xia, Wei Shen, Elliot K. Fishman, and Alan L. Yuille. A 3d coarse-to-fine framework for volumetric medical image segmentation. In Proc. 3DV, 2018.
    Google ScholarLocate open access versionFindings
  • Zhuotun Zhu, Yingda Xia, Lingxi Xie, Elliot K. Fishman, and Alan L. Yuille. Multi-scale coarse-to-fine segmentation for screening pancreatic ductal adenocarcinoma. In Proc. MICCAI, 2019.
    Google ScholarLocate open access versionFindings
Your rating :
0

 

Tags
Comments