Microsoft Word - 4 易建华.doc

Size: px

Start display at page:

Download "Microsoft Word - 4 易建华.doc"

搅太叔
5 years ago
Views:

1 ISSN CODEN JKYTA8 Journal of Frontiers of Computer Science and Technology /011/05(11) Tel: DOI: /j.issn CT * 易建华 1, 何晖光 1+, 赵明昌 1, 1. 中国科学院自动化研究所复杂系统管理与控制国家重点实验室, 北京复旦大学电子工程系, 上海 Coronary Artery Centerline Extraction in Coronary CTA * YI Jianhua 1, HE Huiguang 1+, ZHAO Mingchang 1, 1. State Key Laboratory of Management and Control for Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing , China. Electronic Engineering Department, Fudan University, Shanghai 00433, China + Corresponding author: huiguang.he@ia.ac.cn YI Jianhua, HE Huiguang, ZHAO Mingchang. Coronary artery centerline extraction in coronary CTA. Journal of Frontiers of Computer Science and Technology, 011, 5(11): Abstract: This paper proposes a method based on Fast Marching to extract coronary artery centerline. In the method, firstly, the original image is down-sampled; secondly, the vessel enhancing diffusion filter is used to reduce noise and increase the vessel structure; thirdly, Fast Marching algorithm is used to extract the minimal path of the initial points in down-sampled image; finally, the centerline in the original image is calculated by using minimal path. The paper uses the common data set to evaluate the method. Key words: coronary artery; centerline extraction; coronary CTA 提出了一种冠状动脉中心线的提取方法, 该方法以 Fast Marching( 快速行进 ) 算法为基础, 首先对图像进行下采样, 接着使用血管增强滤波器对图像进行滤波, 然后利用 Fast Marching 算法提取下采样图像中初始点之间的最短路径, 并使用最短路径在原始图像上计算冠状动脉中心线在公共数据集上的验证实验表明, 该方法具有较好的鲁棒性和准确性 *The National Natural Science Foundation of China under Grant No ( ); the National High-Tech Research and Development Plan of China under Grant No. 007AA01Z37 ( (863)). Received 011-0, Accepted

2 1000 Journal of Frontiers of Computer Science and Technology 011, 5(11) 冠状动脉 ; 中心线提取 ; CT 冠脉造影 A TP39 1 CT (coronary CTA), CT CT,,, coronary CTA coronary CTA (level set) Osher Sethian [1], Cohen [] (minimal path) [3], Deschamps [4] Deschamps, Fast Marching coronary CTA,, Fast Marching,,, Fast Marching Osher Sethian 1988, Φt + F Φ = 0 (1), Φ Level Set ; F,, F () ( ) Level Set Fast Marching F>0 F<0,, T, T TF= 1 () Sethian [5], x + x max( D T,0) + min( D T,0) + y + y max( D T,0) + min( D T,0) + (3) z + z 1 max( D T,0) + min( D T,0) = F 3, 1 Fig.1 Coronary artery extraction framework coronary CTA 51

CT 1001 51 51, 1 (voxel), coronary CTA Fast Marching,, 3, [ π/,π/],,, coronary CTA 56 56 56, Fig. Original coronary CTA (No.00) coronary CTA ( 00 ) Fig.

00) 3 coronary CTA ( 00 ) 1998 Frangi Hessian [6],, Manniesing Frangi (vessel enhancing diffusion, VED) [7] (1) P, Hessian I xx Ixy Ixz H( P) = Iyx Iyy Iyz (4) I zx Izy I zz λ 1 λ λ 3 () Vs ( λ) =

3 CT , 1 (voxel), coronary CTA Fast Marching,, 3, [ π/,π/],,, coronary CTA , Fig. Original coronary CTA (No.00) coronary CTA ( 00 ) Fig.3 Fourier transformation of original coronary CTA (No.00) 3 coronary CTA ( 00 ) 1998 Frangi Hessian [6],, Manniesing Frangi (vessel enhancing diffusion, VED) [7] (1) P, Hessian I xx Ixy Ixz H( P) = Iyx Iyy Iyz (4) I zx Izy I zz λ 1 λ λ 3 () Vs ( λ) = 0, λ > 0 or λ3 > 0 (5) R R B S c A β γ λ λ3 (1 e α )e (1 e )e, otherwise λ1 λ, R B =, R A = λλ λ, S = λ1 + λ + λ3 3 3 (3) σ, Vs ( λ ), V = max V s ( λ) σ σ σ min max (6) (4) (diffusion tensor), T D = QΛ Q (7) λ1 0 0, Q Hessian, Λ = 0 λ 0, 0 0 λ 3 1 λ 1 = 1 + ( w 1) V S, λ = λ3 = 1 + ( ε 1) V S (5) D [8] = ( D L) (8) Lt, 4 4, VED,, 3., { } 1 EC ( ) = Ω ω C() s + ω C() s + PCs ( ())ds (9) 1

100 Journal of Frontiers of Computer Science and Technology 011, 5(11) Fig.

4 Vessel enhancing diffusion filter 4, C(s), Ω, L, ω = 0, (1) s, C () s =1, P = P+ ω, (9) : P0, P P0, P { Ω } U( p) = inf

+ { i 1, j, k i+ 1, j, k } { u Ui, j 1, k u Ui, j+ 1, k } { + } (max,,0 ) + (1) (max u U, u U,0 ) = P i, j, k 1 i, j, k 1

4 100 Journal of Frontiers of Computer Science and Technology 011, 5(11) Fig.5 3D reconstruction of minimal path and CTA 5 CT Fig.4 Vessel enhancing diffusion filter 4, C(s), Ω, L, ω = 0, (1) s, C () s =1, P = P+ ω, (9) : P0, P P0, P { Ω } U( p) = inf E( C) = inf P ( C( s)) ds (10) A A, U ( ), P 0 P P 0 P U U Eikonal U = P (11) Deschamps Cohen 000 [4], (max u U, u U,0 ) + { i 1, j, k i+ 1, j, k } { u Ui, j 1, k u Ui, j+ 1, k } { + } (max,,0 ) + (1) (max u U, u U,0 ) = P i, j, k 1 i, j, k 1 i, j (1) U, P P 0, ( 1 ) 5 3.3,,,,,,,, P i, i=,3,,n, N P i (i=,3,,n), P P, P i Cutplane ( i = i i 1,3,..., N) 0.41~0.50 cm, 0.31~0.41 cm, CTA 0.3~0.6 mm 3[9], 17, b = 45, 6, (), Fig.6 Coronary artery cut plane along minimal path 6 3.4,, i

(http://coronary.bigr.nl/index.php) 1 Table 1 Image quality of dataset 1 3 3 8 (right coronary artery, RCA), 3, 51 51 51, 0.3~0.

5 CT 1003, [10],, 7 Fig.7 Segmentation results of coronary artery in cut plane 7, 8, Fig.8 Coronary artery cut plane along centerline 8 4 8, 3D Segmentation in the Clinic: A Grand Challenge II - Coronary Artery Tracking ( 1 Table 1 Image quality of dataset (right coronary artery, RCA), 3, , 0.3~0.6 mm 3 3 (overlap, OV) (average distance, AD) (average distance inside vessel, AI),,,, 1 mm, ;, 1 mm,,,, 5 Fast Marching, coronary CTA, 3D Segmentation in the Clinic: A Grand Challenge II - Coronary Artery Tracking, Rotterdam Coronary

6 1004 Journal of Frontiers of Computer Science and Technology 011, 5(11) Table Accuracy of centerline in RCA (after cut) RCA ( ) OV std/(%) AD std/mm AI std/mm data ± ± ± data ± ± ± data ± ± ± data ± ± ± data ± ± ± data ± ± ± 0.00 data ± ± ± data ± ± ± Artery Algorithm Evaluation Framework coronary CTA, , Fast Marching,,, , ;, ;,,, ;, ;,,, OV 87%~100%, AD 0.5~0.64 mm, AI 0.5~0.60 mm;, References: [1] Osher S, Sethian J A. Fronts propagating with curvature- dependent speed: algorithms based on Hamilton-Jacobi formulations[j]. Journal of Computational Physics, 1988, 79(1): [] Kass M, Witkin A, Terzopoulos D. Snakes: active contour models[j]. International Journal of Computer Vision, 1988, 4(1): [3] Cohen L D, Kimmel R. Global minimum for active contour models: a minimal path approach[j]. International Journal of Computer Vision, 1997, 4(1): [4] Deschamps T, Cohen L D. Minimal paths in 3D images and application to virtual endoscopy[c]//lncs 1843: Proceedings of the 6th European Conference on Computer Vision-Part II (ECCV 00). London, UK: Springer- Verlag, 000: [5] Sethian J A. A fast marching level set method for monotonically advancing fronts[j]. Proceedings of the National Academy of Sciences of the United States of America, 1996, 93(4): [6] Frangi A F, Niessen W J, Vincken K L, et al. Multiscale vessel enhancement filtering[c]//lncs 1496: Proceedigs of the Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI 98). Berlin: Springer-Verlag, 1998: [7] Manniesing R, Viergever M A, Niessen W J. Vessel enhancing diffusion: a scale space representation of vessel structures[j]. Medical Image Analysis, 006, 10(6): [8] Pernoa P, Malik J. Scale-space and edge detection using anisotropic diffusion[j]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1990, 1(7): [9] Wang Wenqing. Pathophysiology of coronary artery anatomy and coronary heart disease[j]. Chinese Journal of Rural Medicine, 003, 1(1): [10] Gonzalez R C, Woods R E. Digital image processing[m]. New Jersey: Prentice-Hall, 00: 附中文参考文献 : [9]. [J]., 003, 1(1): 14 1.

CT 1005 YI Jianhua was born in 1984. He received his M.S. degree from Chinese Academy of Sciences in 010.

degree from Institute of Automation, Chinese Academy of Sciences in 00.

His research interests include medical image processing, pattern recognition and computer graphics, etc.

7 CT 1005 YI Jianhua was born in He received his M.S. degree from Chinese Academy of Sciences in 010. His research interests include medical image processing and pattern recognition, etc. (1984 ),,, 010,, HE Huiguang was born in He received his Ph.D. degree from Institute of Automation, Chinese Academy of Sciences in 00. Now he is an associate professor at Institute of Automation, Chinese Academy of Sciences. His research interests include medical image processing, pattern recognition and computer graphics, etc. He has published more than 70 papers and has been in charge of many national projects. (1973 ),,, 00,,,, 70, 863 ZHAO Mingchang was born in He received his Ph.D. degree from Institute of Automation, Chinese Academy of Sciences in 004. Now he is an associate professor and postdoctoral researcher at Fudan University. His research interests include medical image processing and computer graphics, etc. (1977 ),,, 004,,,,

2 3. 1,,,.,., CAD,,,. : 1) :, 1,,. ; 2) :,, ; 3) :,; 4) : Fig. 1 Flowchart of generation and application of 3D2digital2building 2 :.. 3 : 1) :,

2 3. 1,,,.,., CAD,,,. : 1) :, 1,,. ; 2) :,, ; 3) :,; 4) : Fig. 1 Flowchart of generation and application of 3D2digital2building 2 :.. 3 : 1) :, 3 1 Vol. 3. 1 2008 2 CAA I Transactions on Intelligent Systems Feb. 2008, (,210093) :.,; 3., 3. :; ; ; ; : TP391 :A :167324785 (2008) 0120001208 A system f or automatic generation of 3D building models