Modeling of the human upper airway from multimodal 3d dentofacial images

MODELING OF HUMAN UPPER AIRWAY FROM MULTIMODAL 3D DENTOFACIAL IMAGES BUI NHAT LINH (M.Eng, National University of Singapore) A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING NATIONAL UNIVERSITY OF SINGAPORE 2014 Declaration I hereby declare that this thesis is my original work and it has been written by me in its entirety. I have duly acknowledged all the sources of information which have been used in the thesis. This thesis has also not been submitted for any degree in any university previously. Bui Nhat Linh Sep 2014 ii Acknowledgments I would like to thank my supervisors, Assoc. Prof. Ong Sim Heng and Assoc. Prof. Kelvin Foong Weng Chiong, for their constant guidance and support, without which the work presented in this thesis could not possibly be done. I also express many thanks to all the students in the Vision and Image Processing Laboratory, especially Dr. Hiew Litt Teen, Dr. Nguyen Tan Dat, and Dr. Li Shimao for their advice, discussion, and encouragement. Finally, I would like to thank my wife and my family who always support and encourage me during my candidature. iii Contents Abbreviations xi List of Tables xiii List of Figures xiv Introduction 1.1 Motivation 1.2 Previous work 1.2.1 Segmentation of upper airway from CBCT 1.2.2 Segmentation of upper airway from MR images 1.2.3 Remaining upper airway modeling problem 1.3 Thesis objectives and outline 1.3.1 Objectives 1.3.2 Outline 1.4 Thesis contributions 10 Preliminaries 12 2.1 Upper airway anatomy 12 2.1.1 Pharynx 13 2.1.2 Nose and nasal cavity 14 iv 2.2 Imaging modalities 15 2.2.1 Conebeam CT 16 2.2.2 Magnetic resonance imaging 18 2.2.3 Digitized dental study model 20 2.3 Medical image segmentation 22 2.3.1 Otsu thresholding 22 2.3.2 Morphological processing 23 2.3.3 Active Contour 24 2.3.4 Level set method 26 2.3.5 Graph-cut image segmentation 29 Automatic segmentation of the nasal cavity and paranasal sinuses from cone-beam CT images 33 3.1 Introduction 33 3.2 Materials and Method 36 3.2.1 Materials 36 3.2.2 Method 37 3.3 Experimental results 49 3.3.1 Qualitative results 49 3.3.2 Quantitative result 52 3.4 Discussion 55 3.5 Conclusion 58 Segmentation of thin volume structure: application to the nasal passage in head MRI 59 4.1 Introduction 59 v 4.2 Related work 63 4.2.1 Hessian-based filter 63 4.2.2 Graph-cut segmentation 64 4.3 Materials and method 65 4.3.1 Materials 65 4.3.2 Method 65 4.3.3 Validation 72 4.4 Experimental results 74 4.5 Discussion 78 4.6 Conclusion 81 Registration of MR images and dental surface scans for upper airway modeling 82 5.1 Introduction 82 5.2 Tooth surface model extraction from MR images 85 5.2.1 Anisotropic diffusion 85 5.2.2 Tooth segmentation from MR images 88 5.2.3 Marching cubes surface reconstruction 89 5.2.4 Surface extraction results 91 5.3 Registration of laser surface scan to the reconstructed teeth surface 91 5.3.1 Landmark-based coarse registration 92 5.3.2 Fine registration using ICP 93 5.4 Pharyngeal airway segmentation 5.4.1 Automatic initialization vi 97 97 5.4.2 Level set segmentation 98 5.4.3 Segmentation results 99 5.5 Visualization of the upper airway 100 5.6 Discussion and Concluding Remarks 101 Conclusion and Future Work 6.1 Overview 6.1.1 105 Segmentation of the nasal cavity and paranasal sinuses from CBCT images 6.1.2 106 Segmentation of thin volumetric structure: application to nasal passage in head MRI 6.1.3 105 106 Registration of MR images and dental surface scans for upper airway modeling 6.2 Future work 107 108 References 110 vii Summary A patient-specific virtual upper airway model is important for clinical, education and research applications of the human upper airway such as obstructive sleep apnea (OSA), airflow modeling, and speech production. In this thesis, we present the methods for the segmentation and reconstruction of the human upper airway from multi-modal 3D dentofacial images such as cone-beam computed tomography (CBCT), magnetic resonance (MR) images, and laser surface scan. The nasal cavity and paranasal sinuses are automatically segmented from CBCT images by using novel level set methods. A graph-based segmentation method is developed to segment thin structures from volumetric medical images such as the nasal passage from MR images. A laser surface scan of the dental study model is registered to MR images to visualize the upper airway. We present an automated method for the segmentation of the nasal cavity and paranasal sinuses from CBCT images. Gaussian mixture model thresholding and morphological operators are first employed to automatically locate the region of interest and to initialize the active contour. Second, the active contour driven by the Kullback-Leibler (K-L) divergence energy implemented via the level set is used to segment the upper airway. A new approach is proposed to handle the K-L divergence asymmetry to directly minimize the K-L divergence energy on the probability density function of the image intensity. Finally, to refine the segmentation result, we introduce an anisotropic localized active contour which defines the local area based on shape prior information. Our segmentation method is shown viii to have the capability to delineate the nasal cavity and paranasal sinuses from CBCT images, and have potential for clinical usage. Segmentation results confirm that the proposed method is more accurate than current CBCT segmentation methods such as global or localized region-based level set. We propose a graph-based method for the segmentation of thin volumetric structures such as the nasal passage in MR images. First, a novel sheetness filter based on the eigenvalues of the second order local structure (Hessian) is applied. Second, the medial surface of the structure is estimated by using gradient vector flow. Third, the sheetness measure, medial surface location, and local thickness obtained from the above steps are used as the shape prior in a graph cut method to finally segment the objects. The proposed method is then applied to segment the nasal passage from MR images. Segmentation results demonstrates that the method is more accurate than the min-cut graph cut and the sheetness filter level set method in segmentation of the nasal passage. It is the first study on the nasal passage segmentation from MR images. We develop a method to integrate a laser surface scan of the dental study model and head MR images to extract and visualize the upper airway. The advantage of this approach is only non-radiation imaging modalities are involved. The proposed method consists of the segmentation of the teeth and pharyngeal airway from MR images using level set techniques, and the registration of the laser surface scan to MR images of the head. The reason to register the tooth structures to MR images is that the scanned dental model is superior to MRI in imaging the tooth crown. ix In conclusion, the thesis presents three image processing methods for the modeling of the human upper airway from multimodal 3D dentofacial images. The experiments described in the thesis demonstrate the performance of each method in upper airway segmentation and reconstruction. x study model is registered to MR images to visualize the upper airway. 6.1.1 Segmentation of the nasal cavity and paranasal sinuses from CBCT images We have presented an automated method for the segmentation of the nasal cavity and paranasal sinuses from CBCT images. Gaussian mixture model thresholding and morphological operators are first employed to automatically locate the region of interest and to initialize the active contour. Second, the active contour driven by the K-L divergence energy implemented via the level set is used to segment the upper airway. A new approach is proposed to handle the K-L divergence asymmetry to directly minimize the K-L divergence energy on the probability density function of the image intensity. Finally, to refine the segmentation result, we have introduced an anisotropic localized active contour which defines the local area based on shape prior information. Our segmentation method is shown to have the capability to delineate the nasal cavity and paranasal sinuses from CBCT images, and have potential for clinical usage. Segmentation results confirm that the proposed method is more accurate than current CBCT segmentation methods such as global or localized region-based level set. 6.1.2 Segmentation of thin volumetric structure: application to nasal passage in head MRI We have proposed a graph-based method for the segmentation of thin volumetric structures such as the nasal passage in MR images. First, a novel 106 sheetness filter based on the eigenvalues of the second order local structure (Hessian) is applied. Second, the medial surface of the structure is estimated by using gradient vector flow. Third, the sheetness measure, medial surface location, and local thickness obtained from the above steps are used as the shape prior in a graph cut method to finally segment the objects. The proposed method is then applied to segment the nasal passage from MR images. Segmentation results demonstrates that the method is more accurate than the min-cut graph cut and the sheetness filter level set method in segmentation of the nasal passage. It is the first study on the nasal passage segmentation from MR images. 6.1.3 Registration of MR images and dental surface scans for upper airway modeling We have developed a method to integrate a laser surface scan of the dental study model and head MR images to extract and visualize the upper airway. The advantage of this approach is only non-radiation imaging modalities are involved. The proposed method consists of the segmentation of the teeth and pharyngeal airway from MR images using level set techniques, and the registration of the laser surface scan to MR images of the head. The reason to register the tooth structures to MR images is that the scanned dental model is superior to MRI in imaging the tooth crown. The surface-to-surface registration is the most feasible approach in this application. First, the teeth are segmented from MR images using level set method to build the 3D tooth surface model. A coarse-to-fine registra- 107 tion based on ICP algorithm is applied to align the laser scanned dental study model to the reconstructed tooth surface model. The registered laser scanned dental study model and the reconstructed pharyngeal airway surface model are shown with the MR images in a sectional visualization system. 6.2 Future work Our CBCT segmentation scheme has been tested on limited data sets as CBCT scan does not always include the nasal cavity and paranasal sinus. If the accuracy and reliability of our results can be validated with a larger number of data sets and with CT scan of the same subjects, it can be further evaluated for clinical use. As our method can misclassify the thin bone regions in CBCT images as noise and give an incorrect segmentation, a filter to detect these bones may be incorporated in our segmentation scheme to improve the result. Expanding the proposed scheme to segment other structures in CBCT images such as bone or teeth are also possibilities for future work. Our graph-based method for the segmentation of thin structures from volumetric medical images has only been applied to the nasal passage in MR images of the head. In future, we plan to extend the method to segment other thin volume structures in head MRI such as muscles or bones and also to thin structures in other imaging modalities, such as CBCT (cone beam computed tomography) or CT (computed tomography). The segmentation result of the nasal passage from MRI can be incorrect 108 due to the low contrast or the thin bones which are very close to the nasal passage. The distribution of the sheetness measure can be further modified to be more suitable to the anatomical features of the nasal passage. In our work, user interaction is still necessary for the registration of the laser scanned dental model to MR images of the head. Users have to choose a slice and manual initialize the active contour in this slice for tooth segmentation. To align the two tooth models, corresponding point pairs are manually selected. Automated segmentation of teeth from MR images can be a potential topic for future work. In an other approach, if markers are applied to indicate the corresponding point pairs in MR images and the laser surface scan, the registration can be done with minimum user interaction. 109 References [1] S. M. Bromley, “Smell and taste disorders: a primary care approach,” American Family Physician, vol. 61, pp. 427–436, Jan 2000. [2] W.-H. Jia and H.-D. Qin, “Non-viral environmental risk factors for nasopharyngeal carcinoma: A systematic review,” Seminars in Cancer Biology, vol. 22, no. 2, pp. 117 – 126, 2012. [3] T. Young, P. E. Peppard, and D. J. Gottlieb, “Epidemiology of obstructive sleep apnea: a population health perspective,” American Journal of Respiratory and Critical Care Medicine, vol. 165, pp. 1217–1239, May 2002. [4] M. Lethbridge, J. S. Schiller, and L. Bernadel, “Summary health statistics for U.S. Adults: National health interview survey,” National Center for Health Statistics. Vital Health Statistics, vol. 10, p. 222, 2004. [5] P. Rogalla, “Virtual endoscopy: an application snapshot,” Medica Mundi, vol. 43, no. 1, pp. 17–23, 1999. [6] J. T. M. Boris A. Stuck, “Airway evaluation in obstructive sleep apnea,” Sleep medicine reviews, vol. 12, no. 6, pp. 411–436, 2008. 110 [7] P. Rogalla, A. Nischwitz, A. Heitema, R. Kaschke, and B. Hamm, “Virtual Endoscopy of the Nose and the Paranasal Sinus,” European Radiology, vol. 16, pp. 787–789, 1998. [8] G. S. Ruthenbeck, J. Hobson, A. S. Carney, S. Sloan, R. Sacks, and K. J. Reynolds, “Toward photorealism in endoscopic sinus surgery simulation,” American Journal of Rhinology and Allergy, vol. 27, no. 2, pp. 138–143, 2013. [9] J. Dang and K. Honda, “Construction and control of a physiological articulatory model,” Journal of the Acoustical Society of America, vol. 115, pp. 853–870, Feb 2004. [10] R. Pauwels, J. Beinsberger, B. Collaert, C. Theodorakou, J. Rogers, A. Walker, L. Cockmartin, H. Bosmans, R. Jacobs, R. Bogaerts, and K. Horner, “Effective dose range for dental cone beam computed tomography scanners,” European Journal of Radiology, vol. 81, no. 2, pp. 267 – 271, 2012. [11] M. Loubele, R. Bogaerts, E. V. Dijck, R. Pauwels, S. Vanheusden, P. Suetens, G. Marchal, G. Sanderink, and R. Jacobs, “Comparison between effective radiation dose of CBCT and MSCT scanners for dentomaxillofacial applications,” European Journal of Radiology, vol. 71, no. 3, pp. 461 – 468, 2009. [12] P. Sutthiprapaporn, K. Tanimoto, M. Ohtsuka, T. Nagasaki, M. Konishi, Y. Iida, and A. Katsumata, “Improved inspection of the lateral 111 pharyngeal recess using cone-beam computed tomography in the upright position,” Oral Radiology, vol. 24, pp. 71–75, 2008. [13] A. Katsumata, A. Hirukawa, M. Noujeim, S. Okumura, M. Naitoh, M. Fujishita, E. Ariji, and R. P. Langlais, “Image artifact in dental cone-beam CT,” Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontology, vol. 101, pp. 652–657, 2006. [14] N. A. Alsufyani, C. Flores-Mir, and P. W. Major, “Three-dimensional segmentation of the upper airway using cone beam CT: a systematic review,” Dentomaxillofacial Radiology, vol. 41, pp. 276–284, May 2012. [15] T. Ogawa, R. Enciso, W. H. Shintaku, and G. T. Clark, “Evaluation of cross-section airway configuration of obstructive sleep apnea,” Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontology, vol. 103, pp. 102–108, 2007. [16] H. El and J. M. Palomo, “Measuring the airway in three dimensions: A reliability and accuracy study,” American Journal of Orthodontics and Dentofacial Orthopedics, vol. 137, pp. S50.e1–S50.e9, 2010. [17] M. Celenk, M. L. Farrell, H. Eren, K. Kumar, G. D. Singh, and S. Lozanoff, “Upper airway detection and visualization from cone beam image slices,” Journal of X-Ray Science and Technology, vol. 18, pp. 121–135, 2010. [18] H. Shi, W. C. Scarfe, and A. G. Farman, “Upper airway segmentation and dimensions estimation from cone-beam CT image datasets,” Inter- 112 national Journal of Computer Assisted Radiology and Surgery, vol. 1, pp. 177–186, 2006. [19] I. Cheng, S. Nilufar, C. Flores-Mir, and A. Basu, “Airway segmentation and measurement in CT images,” in Proceeding of 29th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 2007. [20] S. Stratemann, J. C. Huang, K. Maki, D. Hatcher, and A. J. Millere, “Three-dimensional analysis of the airway with cone-beam computed tomography,” American Journal of Orthodontics and Dentofacial Orthopedics, vol. 140, pp. 607–615, 2011. [21] J. Liu, J. K. Udupa, D. Odhnera, J. M. McDonough, and R. Arens, “System for upper airway segmentation and measurement with MR imaging and fuzzy connectedness,” Academic Radiology, vol. 10, pp. 13–24, Jan 2003. [22] M. B. Abbott, B. J. Dardzinski, and L. F. Donnelly, “Using volume segmentation of cine MR data to evaluate dynamic motion of the airway in pediatric patients,” American Journal of Roentgenol, vol. 181, pp. 857–859, Sep 2003. [23] K. C. Welch, G. D. Foster, C. T. Ritter, T. A. Wadden, R. Arens, G. Maislin, and R. J. Schwab, “A novel volumetric magnetic resonance imaging paradigm to study upper airway anatomy,” Sleep, vol. 25, pp. 532–542, Aug 2002. 113 [24] J. Behrends and A. Wismuller, “A segmentation and analysis method for MRI data of the human vocal tract,” in Proceedings of the Symposium on Human and Machine Perception in Acoustic and Visual Communication, 2001. [25] S. Ventura, D. Freitas, and J. Tavares, “Imaging of the vocal tract based on magnetic resonance techniques,” in Computer Vision, Imaging and Computer Graphics: Theory and Applications, vol. 68, pp. 146– 157, 2010. [26] E. Bresch and S. Narayanan, “Region segmentation in the frequency domain applied to upper airway real-time magnetic resonance images,” IEEE Transactions on Medical Imaging, vol. 28, pp. 323–338, Mar 2009. [27] L. A. Feldkamp, L. C. Davis, and J. W. Kress, “Practical cone-beam algorithm,” Journal of the Optical Society of American A, vol. 1, pp. 612–619, Jun 1984. [28] S. Adibi, W. Zhang, T. Servos, and P. N. O’Neill, “Cone beam computed tomography in dentistry: what dental educators and learners should know,” Journal of Dental Education, vol. 76, pp. 1437–1442, Nov 2012. [29] N. Otsu, “A threshold selection method from grey-level histograms,” IEEE Transactions on Systems, Man and Cybernetics, vol. 9, pp. 41– 47, Jan 1979. 114 [30] J. P. Serra, Image Analysis and Mathematical Morphology. Academic Press Inc., 1982. [31] M. Kass, A. Witkin, and D. Terzopoulos, “Snakes: Active contour models,” International Journal of Computer Vision, vol. 1, no. 4, pp. 321–331, 1988. [32] J. Sethian, Level set methods and fast marching methods. 2001. [33] V. Caselles, R. Kimmel, and G. Sapiro, “Geodesic active contours,” International Journal of Computer Vision, vol. 22, no. 1, pp. 61–97, 1997. [34] T. F. Chan and L. A. Vese, “Active contours without edges,” IEEE Transactions on Image Processing, vol. 10, no. 2, pp. 266–277, 2001. [35] O. Michailovich, Y. Rathi, and A. Tannenbaum, “Image segmentation using active contours driven by the bhattacharyya gradient flow,” IEEE Transactions on Image Processing, vol. 16, pp. 2787–2801, Nov 2007. [36] L. R. Ford and D. R. Fulkerson, “Maximal flow through a network,” Canadian Journal of Mathematics, vol. 8, p. 399, 1956. [37] Y. Boykov and V. Kolmogorov, “An Experimental Comparison of MinCut/Max-Flow Algorithms for Energy Minimization in Vision,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 26, pp. 359–374, 2001. 115 [38] Y. Boykov and M.-P. Jolly, “Interactive graph cuts for optimal boundary region segmentation of objects in N-D images,” in Proceedings of Eighth IEEE International Conference on Computer Vision 2001, vol. 1, pp. 105–112 vol.1, 2001. [39] R. Guijarro-Martinez and G. R. Swennen, “Cone-beam computerized tomography imaging and analysis of the upper airway: a systematic review of the literature,” International Journal of Oral and Maxillofacial Surgery, vol. 40, pp. 1227–1237, Nov 2011. [40] A. Weissheimer, L. M. de Menezes, G. T. Sameshima, R. Enciso, J. Pham, and D. Grauer, “Imaging software accuracy for 3-dimensional analysis of the upper airway,” American Journal of Orthodontics and Dentofacial Orthopedics, vol. 142, no. 6, pp. 801 – 813, 2012. [41] S. Osher and R. Fedkiw, Level set methods and dynamic implicit surfaces. Springer, 2003. [42] S. Kullback and R. A. Leibler, “On information and sufficiency,” The Annals of Mathematical Statistics, vol. 22, pp. 79–86, 03 1951. [43] S. Lankton and A. Tannenbaum, “Localizing region-based active contours,” IEEE Transactions on Image Processing, vol. 17, pp. 1–11, Nov. 2008. [44] A. Myronenko and X. Song, “Global active contour-based image segmentation via probability alignment,” in Proceeding of IEEE Conference on Computer Vision and Pattern Recognition, pp. 2798–2804, June 2009. 116 [45] N. Houhou, J. Thiran, and X. Bresson, “Fast texture segmentation model based on the shape operator and active contour,” in Proceeding of IEEE Conference on Computer Vision and Pattern Recognition, pp. 1–8, June 2008. [46] Q. Zheng, Z. Lu, W. Yang, M. Zhang, Q. Feng, and W. Chen, “A robust medical image segmentation method using Kullback-Leibler distance and local neighborhood information,” Computers in Biology and Medicine, vol. 43, no. 5, pp. 459 – 470, 2013. [47] W. E. Lorensen and H. E. Cline, “Marching cubes: A high resolution 3D surface construction algorithm,” Computer Graphics, vol. 21, pp. 163–169, July 1987. [48] F. Schulze, K. Buhler, A. Neubauer, A. Kanitsar, L. Holton, and S. Wolfsberger, “Intra-operative virtual endoscopy for image guided endonasal transsphenoidal pituitary surgery,” International Journal of Computer Assisted Radiology and Surgery, vol. 5, no. 2, pp. 143– 154, 2010. [49] V. Trevillot, R. Sobral, E. Dombre, P. Poignet, B. Herman, and L. Crampette, “Innovative endoscopic sino-nasal and anterior skull base robotics,” International Journal of Computer Assisted Radiology and Surgery, vol. 8, no. 6, pp. 977–987, 2013. [50] C. F. Westin, A. Bhalerao, H. Knutsson, and R. Kikinis, “Using local 3D structure for segmentation of bone from computer tomography 117 images,” in Proceeding of IEEE Conference on Computer Vision and Pattern Recognition, pp. 794–800, Jun 1997. [51] M. Descoteaux, M. Audette, K. Chinzei, and K. Siddiqi, “Bone enhancement filtering: Application to sinus bone segmentation and simulation of pituitary surgery,” in Proceeding of International Conference on Medical Image Computing and Computer-Assisted Intervention, pp. 9–16, 2005. [52] A. Frangi, W. Niessen, K. Vincken, and M. Viergever, “Multiscale vessel enhancement filtering,” in Proceeding of International Conference on Medical Image Computing and Computer-Assisted Interventation, pp. 130–137, 1998. [53] M. Krcah, G. Szekely, and R. Blanc, “Fully automatic and fast segmentation of the femur bone from 3D-CT images with no shape prior,” in Proceeding of IEEE International Symposium on Biomedical Imaging: From Nano to Macro, pp. 2087–2090, March 2011. [54] C. Reinbacher, T. Pock, C. Bauer, and H. Bischof, “Variational segmentation of elongated volumetric structures,” in Proceeding of IEEE Conference on Computer Vision and Pattern Recognition, pp. 3177– 3184, June 2010. [55] P. Das, O. Veksler, V. Zavadsky, and Y. Boykov, “Semiautomatic segmentation with compact shape prior,” Image and Vision Computing, vol. 27, no. 12, pp. 206 – 219, 2009. 118 [56] G. Slabaugh and G. Unal, “Graph cuts segmentation using an elliptical shape prior,” in Proceeding of IEEE International Conference on Image Processing, vol. 2, pp. II–1222–5, Sept 2005. [57] O. Veksler, “Star Shape Prior for Graph-Cut Image Segmentation,” in Proceeding of European Conference on Computer Vision, pp. III: 454–467, 2008. [58] S. Vicente, V. Kolmogorov, and C. Rother, “Graph cut based image segmentation with connectivity priors,” in Proceeding of IEEE Conference on Computer Vision and Pattern Recognition, pp. 1–8, 2008. [59] S. Esneault, C. Lafon, and J.-L. Dillenseger, “Liver vessels segmentation using a hybrid geometrical moments/graph cuts method,” IEEE Transactions on Biomedical Engineering, vol. 57, pp. 276–283, Feb 2010. [60] C. Bauer, T. Pock, E. Sorantin, H. Bischof, and R. Beichel, “Segmentation of interwoven 3D tubular tree structures utilizing shape priors and graph cuts,” Medical Image Analysis, vol. 14, no. 2, pp. 172 – 184, 2010. [61] N. Zhu and A. Chung, “Optimal and efficient segmentation for 3D vascular forest structure with graph cuts,” in Proceeding of IEEE International Conference on Image Processing, pp. 1135–1139, Sept 2013. [62] C. Xiao, M. Staring, D. Shamonin, J. H. Reiber, J. Stolk, and B. C. Stoel, “A strain energy filter for 3D vessel enhancement with applica- 119 tion to pulmonary CT images,” Medical Image Analysis, vol. 15, no. 1, pp. 112 – 124, 2011. [63] C. Xu and J. L. Prince, “Snakes, shapes, and gradient vector flow,” IEEE Transactions on Image Processing, vol. 7, no. 3, pp. 359–369, 1998. [64] M. Hassouna and A. Farag, “Variational curve skeletons using gradient vector flow,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 31, pp. 2257–2274, Dec 2009. [65] V. Kolmogorov and Y. Boykov, “What Metrics Can Be Approximated by Geo-Cuts, Or Global Optimization of Length/Area and Flux,” in Proceeding of International Conference on Computer Vision, pp. I: 564–571, 2005. [66] A. Oksenberg and E. Arons, “Sleep bruxism related to obstructive sleep apnea: the effect of continuous positive airway pressure,” Sleep Medicine, vol. 3, pp. 513–515, Nov 2002. [67] P. Perona and J. Malik, “Scale-space and edge detection using anisotropic diffusion,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 12, pp. 629–639, Jul 1990. [68] H. Gao and O. Chae, “Individual tooth segmentation from CT images using level set method with shape and intensity prior,” Pattern Recognition, vol. 43, no. 7, pp. 2406 – 2417, 2010. 120 [69] P. Besl and N. D. McKay, “A method for registration of 3-D shapes,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 14, pp. 239–256, Feb 1992. [70] B. Green, “The orthogonal approximation of an oblique structure in factor analysis,” Psychometrika, vol. 17, no. 4, pp. 429–440, 1952. 121 [...]... model is then registered to the tooth surfaces The pharyngeal airway is also segmented to reconstruct the surface model 9 • Chapter 6: This chapter summarizes the results, concludes the thesis, and suggests future research 1.4 Thesis contributions The main contributions of the thesis are the algorithms for the segmentation and reconstruction of the human upper airway from multi-modal 3D dentofacial images. .. upper airway, which is the region of interest of this thesis We discuss the imaging modalities used in our study in Section 2.2 Finally, we give a review of related segmentation methods in Section 2.3 2.1 Upper airway anatomy In this thesis, we aim to develop algorithms for the segmentation of the human upper airway In this section, we describe the anatomy of the human upper airway in detail The human upper. .. method for the segmentation of the thin and elongated structures from volumetric medical images The proposed method is applied to segment the nasal passage from MR images of the head • Chapter 5: This chapter describes an approach to integrate the laser scanned surface dental model to the MR image of the head for modeling the upper airway First the tooth surfaces are extracted from MR images The scanned... part of the upper airway, all the above works did not deal with the segmentation problem of the nasal passage from MR images 7 1.2.3 Remaining upper airway modeling problem While many methods have been proposed for the segmentation of the upper airway from CBCT images, the segmentation of the nasal cavity and para nasal sinuses has not been addressed Similarly, no one has proposed a method for the segmentation... measure If the registration error is out of range, there is no color Out of range happens at the gingiva area on the laser surface scan as the reconstructed tooth surface does not include gingiva 5.7 96 Segmentation of the pharyngeal airway from MR images using 99 level set method 5.8 Segmentation of the pharyngeal airway from MR images 5.9 100 Sectional visualization of the upper airway from MRI and... patientspecific virtual model of upper airway is useful for users to study the human upper airway in clinical, education and research applications Our novel algorithms allow users to examine the complex anatomy of human upper airway such as the nasal cavity and paranasal sinuses from CBCT, nasal passage from MRI, and the teeth from MRI and laser surface scan The significant contributions of the thesis are described... medical images and apply it to segment the nasal passage from MR images of the head • To register the digitized dental study model to MR images of the head so as to visualize the entire upper airway using non-radiation imaging modalities 8 1.3.2 Outline The thesis is divided into six chapters The remaining chapters of the thesis are organized as follows: • Chapter 2: This chapter presents the background of. .. pharynx is the nasopharyngeal airway It is behind the nose, and connected to the nasal cavity The lateral pharyngeal recess, where most NPCs 13 originate, is at the back of the nasopharyngeal airway The oropharyngeal airway is at the bottom of the oral cavity, separated from the nasopharyngeal airway by the palate The oropharyngeal airway is important for OSA studies Figure 2.2: Illustration of Pharynx... cavity The nose and nasal cavity are the main opening of the airway to outside The nose is a structure of soft and hard tissues covering the front part of the nasal cavity The nasal cavity is the empty space above and behind the nose The air is warmed, humidified, and filtered when passing through the nasal cavity The nasal cavity is divided into two by the nasal septum In each side of the nasal cavity, there... segmentation of the nasal passage from MR images Solving the above two remaining problems are the main objectives of this thesis 1.3 1.3.1 Thesis objectives and outline Objectives The aims of our research are: • To develop an automated method for the segmentation of the nasal cavity and paranasal sinuses from CBCT images of the head • To develop a graph-based method to extract thin volumetric structures from 3D . in imaging the tooth crown. ix In conclusion, the thesis presents three image processing methods for the modeling of the human upper airway from multimodal 3D dentofacial images. The experiments. of the segmentation of the upper airway are reviewed. The objectives and outline of the thesis are then presented and this is followed by the contributions of the thesis. 1.1 Motivation The human. MODELING OF HUMAN UPPER AIRWAY FROM MULTIMODAL 3D DENTOFACIAL IMAGES BUI NHAT LINH (M.Eng, National University of Singapore) A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT