采用核磁共振图像的心脏2D运动分析新方法
|
摘要
医学成像术成为心脏疾病诊断的重要手段，核磁共振成像(Magnetic ResonanceImage)是一种用于肌体组织运动分析的理想成像模式，应用于心脏的运动分析已有20多年的历史。基于核磁共振成像的心脏运动分析方法大量涌现，论文主要对采用核磁共振成像的心脏运动进行研究，并提出一种用于心脏运动分析的新方法。主要成果有：
(1)通过对大量有关可控滤波器文献的阅读和研究，总结了可控滤波器的理论及应用。可控滤波器在图像运动分析方面的应用具有重要意义，将图像序列在频域中进行频谱分析，其运动速度可用角度表示，这是使用可控滤波器分析图像运动的理论基础，但是仅适用于简单的平移运动。本文提出改进方法，以图像序列中一个像素点为中心，和它周围较小邻域的像素点组成小块图像序列，将该图像序列与可控滤波器卷积进行分析。实验结果表明，改进的方法能够较好的描述复杂的图像运动。
(2)通过对大量MRI心脏运动分析方法的研究，以及对几种经典方法的优点和弊端的总结，提出了一种用于MRI心脏运动分析的新方法，该方法利用可控滤波器描述心脏图像序列的2D运动。由于心脏运动的复杂性，本文利用可控滤波器改进的运动分析方法求取MRI心脏的运动矢量。实验结果表明，可控滤波器能够较好的描述心脏的2D运动。
Medical imaging becomes an important mean of heart disease diagnosis. Magneticresonance image is a kind of ideal imaging pattern for organism tissues. It has been used toanalyze cardiac motion for more than 20 years. There spring up mess of methods to cardiacMRI motion estimation. In this thesis, we do the research mainly on the cardiac MRI motionand a new method to the cardiac motion estimation is proposed. The main work of this thesisis as follows:
(1) The theory and application of the steerable filter are summarized through the researchof a large number of literatures. The application of motion analysis of steerable filters hasimportant meaning. If the image sequence is analyzed in the frequency domain, its speed canbe represented by the orientation. These are the preconditions of analyzing the motion ofimages with steerable filters. But this method is just suit for simple translation movement. Thepaper puts forward a new method; we take a pixel in the image sequence as the center andwith its surrounding pixels to form a small image sequence. Convolve the image sequencewith steerable filters to analyze. The experimental results show that the improved method candescribe complex motion of images well.
(2) A new method to the cardiac motion estimation is proposed through widely researchon the cardiac motion methods and summarize of the advantages and disadvantages of severalclassical methods. The method is to describe the cardiac 2D motion using the steerable filter.The motion of heart is complex, so in the paper we use the improved method to describe themotion of cardiac MRI, namely, we calculate the velocity of motion of each pixel one by one.The experimental results show that we can get a good result when we use the steerable filtersto the cardiac motion analysis.
(1)通过对大量有关可控滤波器文献的阅读和研究，总结了可控滤波器的理论及应用。可控滤波器在图像运动分析方面的应用具有重要意义，将图像序列在频域中进行频谱分析，其运动速度可用角度表示，这是使用可控滤波器分析图像运动的理论基础，但是仅适用于简单的平移运动。本文提出改进方法，以图像序列中一个像素点为中心，和它周围较小邻域的像素点组成小块图像序列，将该图像序列与可控滤波器卷积进行分析。实验结果表明，改进的方法能够较好的描述复杂的图像运动。
(2)通过对大量MRI心脏运动分析方法的研究，以及对几种经典方法的优点和弊端的总结，提出了一种用于MRI心脏运动分析的新方法，该方法利用可控滤波器描述心脏图像序列的2D运动。由于心脏运动的复杂性，本文利用可控滤波器改进的运动分析方法求取MRI心脏的运动矢量。实验结果表明，可控滤波器能够较好的描述心脏的2D运动。
Medical imaging becomes an important mean of heart disease diagnosis. Magneticresonance image is a kind of ideal imaging pattern for organism tissues. It has been used toanalyze cardiac motion for more than 20 years. There spring up mess of methods to cardiacMRI motion estimation. In this thesis, we do the research mainly on the cardiac MRI motionand a new method to the cardiac motion estimation is proposed. The main work of this thesisis as follows:
(1) The theory and application of the steerable filter are summarized through the researchof a large number of literatures. The application of motion analysis of steerable filters hasimportant meaning. If the image sequence is analyzed in the frequency domain, its speed canbe represented by the orientation. These are the preconditions of analyzing the motion ofimages with steerable filters. But this method is just suit for simple translation movement. Thepaper puts forward a new method; we take a pixel in the image sequence as the center andwith its surrounding pixels to form a small image sequence. Convolve the image sequencewith steerable filters to analyze. The experimental results show that the improved method candescribe complex motion of images well.
(2) A new method to the cardiac motion estimation is proposed through widely researchon the cardiac motion methods and summarize of the advantages and disadvantages of severalclassical methods. The method is to describe the cardiac 2D motion using the steerable filter.The motion of heart is complex, so in the paper we use the improved method to describe themotion of cardiac MRI, namely, we calculate the velocity of motion of each pixel one by one.The experimental results show that we can get a good result when we use the steerable filtersto the cardiac motion analysis.
引文
[1] M. A. Guttman, E. A. Zerhouni. Analysis and visualization of cardiac function from MRimages [J]. IEEE Computer Graphics and Applications, 1997, 17(1): 30~38.
[2] A. F. Frangi, W. J. Niessen, M. A. Viergever. Three-dimensional modeling for functionalanalysis of cardiac images: A review [J]. IEEE Transacions Medical Imaging,2001, 20(1): 2~25.
[3]吉强.医学成像技术发展[J].中华现代影像学杂志2008,4(6).
[4] E. A. Zerhouni, D. Parish, W. Rogers, et al. Human heart: Tagging with MR imaging - amethod for noninvasive assessment of myocardial motion [J]. Radiology, 1988,169:59~63.
[5] L. Axel, L. Dougherty. MR imaging of motion with spatial modulation of magnetization [J].Radiology, 1989,171: 841~845.
[6] L. Axel, L. Dougherty. Heart wall motion: Improved method of spatial modulation ofmagnetization [J]. Radiology, 1989,172:349~350.
[7]王元全，周则明，孙越泓等.带标记线的心脏核磁共振图像分析综述[J].中国图形图象学报，2003, 8(11):1233~1241.
[8] J. L. Prince, ER McVeigh. Motion estimation from tagged MR image sequences[J]. IEEEtrans Med Imag, 1992, 11(2):238~249.
[9] T. S. Denney, JL Prince. Optimal brightness function for optical flow estimation of deformablemotion [J]. IEEE trans IP, 1994, 3(2):178~190.
[10] T. S. Denney, J. L. Prince. Reconstruction of 3D left ventricular motion from planar taggedcardiac MR images: an estimation theoretic approach [J]. IEEE Transactions Medical Imaging, 1995,14(4):625~635.
[11] L. Yan, T. S. Denney. Unsupervised estimation of left ventricular displacement from MRtagged images using Markov Random field edge Priors[J]. International Conference on ImageProcessing, 1998, 1:689~693.
[12] D. Terzopoulos, A. Witkin, M. Kass. Constraints on deformable models: recovering 3Dshape and nonrigid motion[J]. Artificial Inteligence, 1988, 36(1): 91~123.
[13] D. Terzopoulos, A. Witkin. Deformable models [J]. IEEE Computer Graphics andApplications, 1988, 8(1): 41~51.
[14] T. McInerney, D. Terzopoulos. Deformable models in medical image analysis: A survey[J].Medical Image Analysis, 1996, 1(2): 91~108.
[15] T. McInemey, D. Terzopoulos. A dynamic finite element surface model for segmentation andtracking in multidimensional medical images with application to cardiac 4D images analysis [J].Computerized Medical Imaging and Graphics, 1995, 19(1): 69~83.
[16] J. Park, D. Metaxas, A. Young, et al. Deformable models with parameters functions forcardiac motion analysis form tagged MRI data[J]. IEEE Transactions Medical Imaging, 1996, 15(3):278~289.
[17] J. Park, D. Metaxas, L. Axel. Analysis of left-ventricular wall motion based on volumetricdeformable models and MRI-SPAMM [J]. Medical Image Analysis, 1996,1(1): 53~71.
[18] D. Metaxas, D. Terzopoulos. Shape and nonrigid motion estimation through physics~basedsynthesis [J]. IEEE Transactions Pattern Analysis and Machine Intelligence, 1993, 15(6): 569~579.
[19] T. O’Donnel, T. Boult, A. Gupta. Global models with parametric offsets as applied to cardiacmotion recovery [J]. IEEE Computer Society Conference on Computer Vision and PatternRecognition,1996: 293~299.
[20] A. Young, L. D. Kraitchman, L. Dougherty, et al. Tracking and finite element analysis ofstripe deformation in magnetic resonance tagging [J]. IEEE Transactions Medical Imaging, 1995,14(3): 413~421.
[21] A. Young, L. Axel. Three-dimensional motion and deformation in the heart wall: Estimationfrom spatial modulation of magnetization-A model based approach. Radiology, 1992(185): 241~247.
[22] A. Young. Model tags: Direct three-dimensional tracking of heart wall motion from taggedmagnetic resonance images [J]. Media, 1999, 3(4): 361~372.
[23] G. Farin. Curves and surfaces for computer aided geometric design: A practical guide [M].New York: Academic, 1990.
[24] J. M. Moulton. Spline surface interpolation for calculating three-dimensional ventricularstrains from MRI tissue tagging [J]. Physiology, 1996(270): 281~297.
[25] A. A. Amini. Energy-minimizing deformable grids for tracking tagged MR cardiac images
[A]. In: Computer in Cardiology [C]. Durham, North Carolina,USA, 1992：651~654.
[26] A. A. Amini, Y. Chen, W. R. Curwen, et al. Coupled B-snake grids and constrained thin-platesplines for analysis of 2D tissue deformations from tagged MRI [J]. IEEE Transactions MedicalImaging, 1998, 17(3): 344~356.
[27] J. Huang, D. Abendschein, G. V. Davila-Roman, et al. Spatial-Temporal tracking ofmyocardial deformations with a 4D B-spline model from tagged MRI [J]. IEEE Transactions MedicalImaging, 1999, 18(10): 957~972.
[28] C. Ozturk, R. E. McVeigh. Four-dimensional B-spline based motion analysis of taggedcardiac MR images: introduction and in vivo vallication [J]. Physics in Medicine and Biology, 2000,45(6):1683~1702.
[29] J. Declerck, J. Feldmar, N. Ayache. Definition of a 4D continuous planispherictransformation for the tracking and the analysis of LV motion [J]. Medical Image Analysis, 1998, 4(1):1~17.
[30] N. F. Osman, W. S. Kerwin, E.R. McVergh, et al. Cardiac motion tracking using CINEharmonic phase magnetic resonance imaging [J]. Magnetic Resonance in Medicine, 1999, 42(3):1048~1060.
[31] N. F. Osman, J. Prince. Visualizing mylcardial function using HARP MRI [J]. Physics inMedicine and Biology, 2000, 45(6):1665~1682.
[32] N. F. Osman, E. R. McVeigh, J. Prince. Imaging heart motion using harmonic phase MRI[J].IEEE Transactions Medical Imaging, 2000, 19(3): 186~202.
[33] N. F. Osman, J. L. Prince. Regenerating MR tagged images using harmonic phase (HARP)method [J]. IEEE Transactions on Biomedical Engineering, 2004, 51(8): 1428~1433.
[34] L. Pan, J. L. Prince, et al, Fast tracking of cardiac motion using 3D-HARP. IEEE transactionson Biomedical Engineering, 2005, 52(8):1424~1435.
[35] D. Gabor. Theory of communication [J]. Journal of the IEEE, 1946, 9(3): 429~457.
[36] W. T. Freeman, E. H. Adelson. The design and use of steerable filters[J]. IEEE Transactionson Pattern Analysis and Machine Intelligence, 1991, 13(9):891~906.
[37] C. L. Huang, Y. T. Chen. Motion estimation method using a 3D steerable filter[J]. Image andVision Computing, 1995, 13(1):.21~32.
[38] E. H. Adelson, J. R. Bergen. Spatiotemporal energy models for the perception of motion[J]. J.Opt. Soc. Am., 1995, 2(2): 284~299.
[39] E. P. Simoncelli, H. Farid. Steerable Wedge Filters for Local Orientation Analysis[J]. IEEETransactions on Image Processing, 1996, 5(9): 1377~1382.
[40] W. Yu, K. Daniilidis, G. Sommer. A new 3D orientation steerale filter[J]. In Proceedings ofDAGM-Symposium, 2000:203~212.
[41] J. M. Gryn. Three-dimensional nth derivative of gaussian separable steerable filters[J]. IEEEInternational Conference on Image Processing, 2005, 3:III–553-6.
[42] S. Xu, Y. Jia, X. Zhang. Coding facial expression with oriented steerable filters[J]. ICIP,2006: 2057~2060.
[43] N. Sahba, V. Tavakoli. Ultrasound thermal change detection based on steerable filters. 30thAnnual International IEEE EMBS Conference Vancouver, British Columbia, Canada, 2008:2984~2987.
[44] M. Muehlich, D. Friedrich, T. Aach. Design and implementation of Multi-Steerable MatchedFilters[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011.
[45] K. Krajsek, R. Mester. A unified theory for steerable and quadrature filters. Communicationsin Computer and Information Science, 2007, 4(10):.201~214.
[46] W. Zhou, H. Shu. Detection of Cerebral Vessels in MRA Using 3D Steerable Filters [J].Engineering in Medicine and Biology Society, 2005: 3249~3252.
[47] E. P. Simoncelli. The steerable pyramid: a flexible architecture for multi-scale derivativecomputation[C]. 2ndAnnual IEEE Conference on Image Processing, 1995,3: 444~447.
[48]顾宇灏，林青，胡波.一种基于方向可控金字塔结构的快速医学图像配准算法[J].复旦学报, 2010,49(2): 177~184.
[49]王蕊，尹忠科，龙奕.基于可控金字塔的轮廓波变换构造及其应用[J].铁道学报，2009,31(6): 53~57.
[50] X. Wang, D. D. Feng. Medical image registration via steerable pyramid[C]. Proceeding ofthe 29thAnnual International Conference of the IEEE EMBS, 2007: 6395~6398.
[51]顾立娟,邵命山,郝玉保.基于可控金字塔子带能量特征的文种识别方法[J].计算机应用与软件, 2011, 28(3):91~94.
[52]何青.利用方向可控金字塔进行基于互信息的图像配准[J].计算机应用, 2005, 25(12):2834~2836.
[53] B.K.P. Horn, B.G. Schunck. Determining optical flow [J]. Artificial Intelligence, 1981,17:185~204.
[54] M. Turner. Texture discrimination by gabor function [J]. Biological Cybernetics,1986,55:.71~82.
[55] Y. Zhu, T. Tan, Y. Wang. Biometric personal identification based on iris patterns [J].Proceedings of 15thInternational Conference on Pattern Recognition, 2000:801~804.
[56] C. Liu, H. Wechsler. Gabor feature based classification using the enhanced flasher linersdiscriminate model for face recognition[J]. IEEE Transactions on Image Processing, 2003, 11(2):.467~476.
[57] L. Hong, Y. Wan, A. Jain. Fingerprint image enhancement: algorithm and performanceevaluation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1998, 20(8): 777~789.
[58] A. Jain, K. N. Ratah, S. Lakshmanan. Object detection using gaborfilters[J]. PatternRecognition, 1997, 30(2): 295~309.
[59] D. N. Metaxas, L. Axel, Z. Qian, et al. A segmentation and Tracking System for 4D CardiacTagged MR Images[J]. IEEE EMBS Annual International Conference, 2006:1541~1544.
[60] L. Axel, S. Chung, T. Chen. Tagged MRI Analysis Using Gabor Filters[J]. IEEEInternational Symposium on Biomedical Imaging, 2007: 684~687.
[61] T. Chen, S. Chung, L. Axel. Automated Tag Tracking Using Gabor Filter Bank, Robust PointMatching, and Deformable Models[J]. FIMH, 2007: 22~31.
[62] T. Chen, X. Wang, D. Metaxas, et al. 3D Cardiac Motion Tracking Using Robust PointMatching and Models[J]. IEEE International Symposium on Biomedical Imaging, 2008:280~283.
[63] J. J. Koenderink, A. J. van Doorn. Representation of local geometry in the visual system[J].Biol. Cybern., 1987, 55: 367~375.
[64] H. Knutsson, R. Wilson, G. H. Granlund. Anisotropic non-stationary image estimation and itsapplications: Part 1-Restoration of noisy images[J]. IEEE Trans. Commun., 1983, 31(3):.388~397.
[2] A. F. Frangi, W. J. Niessen, M. A. Viergever. Three-dimensional modeling for functionalanalysis of cardiac images: A review [J]. IEEE Transacions Medical Imaging,2001, 20(1): 2~25.
[3]吉强.医学成像技术发展[J].中华现代影像学杂志2008,4(6).
[4] E. A. Zerhouni, D. Parish, W. Rogers, et al. Human heart: Tagging with MR imaging - amethod for noninvasive assessment of myocardial motion [J]. Radiology, 1988,169:59~63.
[5] L. Axel, L. Dougherty. MR imaging of motion with spatial modulation of magnetization [J].Radiology, 1989,171: 841~845.
[6] L. Axel, L. Dougherty. Heart wall motion: Improved method of spatial modulation ofmagnetization [J]. Radiology, 1989,172:349~350.
[7]王元全，周则明，孙越泓等.带标记线的心脏核磁共振图像分析综述[J].中国图形图象学报，2003, 8(11):1233~1241.
[8] J. L. Prince, ER McVeigh. Motion estimation from tagged MR image sequences[J]. IEEEtrans Med Imag, 1992, 11(2):238~249.
[9] T. S. Denney, JL Prince. Optimal brightness function for optical flow estimation of deformablemotion [J]. IEEE trans IP, 1994, 3(2):178~190.
[10] T. S. Denney, J. L. Prince. Reconstruction of 3D left ventricular motion from planar taggedcardiac MR images: an estimation theoretic approach [J]. IEEE Transactions Medical Imaging, 1995,14(4):625~635.
[11] L. Yan, T. S. Denney. Unsupervised estimation of left ventricular displacement from MRtagged images using Markov Random field edge Priors[J]. International Conference on ImageProcessing, 1998, 1:689~693.
[12] D. Terzopoulos, A. Witkin, M. Kass. Constraints on deformable models: recovering 3Dshape and nonrigid motion[J]. Artificial Inteligence, 1988, 36(1): 91~123.
[13] D. Terzopoulos, A. Witkin. Deformable models [J]. IEEE Computer Graphics andApplications, 1988, 8(1): 41~51.
[14] T. McInerney, D. Terzopoulos. Deformable models in medical image analysis: A survey[J].Medical Image Analysis, 1996, 1(2): 91~108.
[15] T. McInemey, D. Terzopoulos. A dynamic finite element surface model for segmentation andtracking in multidimensional medical images with application to cardiac 4D images analysis [J].Computerized Medical Imaging and Graphics, 1995, 19(1): 69~83.
[16] J. Park, D. Metaxas, A. Young, et al. Deformable models with parameters functions forcardiac motion analysis form tagged MRI data[J]. IEEE Transactions Medical Imaging, 1996, 15(3):278~289.
[17] J. Park, D. Metaxas, L. Axel. Analysis of left-ventricular wall motion based on volumetricdeformable models and MRI-SPAMM [J]. Medical Image Analysis, 1996,1(1): 53~71.
[18] D. Metaxas, D. Terzopoulos. Shape and nonrigid motion estimation through physics~basedsynthesis [J]. IEEE Transactions Pattern Analysis and Machine Intelligence, 1993, 15(6): 569~579.
[19] T. O’Donnel, T. Boult, A. Gupta. Global models with parametric offsets as applied to cardiacmotion recovery [J]. IEEE Computer Society Conference on Computer Vision and PatternRecognition,1996: 293~299.
[20] A. Young, L. D. Kraitchman, L. Dougherty, et al. Tracking and finite element analysis ofstripe deformation in magnetic resonance tagging [J]. IEEE Transactions Medical Imaging, 1995,14(3): 413~421.
[21] A. Young, L. Axel. Three-dimensional motion and deformation in the heart wall: Estimationfrom spatial modulation of magnetization-A model based approach. Radiology, 1992(185): 241~247.
[22] A. Young. Model tags: Direct three-dimensional tracking of heart wall motion from taggedmagnetic resonance images [J]. Media, 1999, 3(4): 361~372.
[23] G. Farin. Curves and surfaces for computer aided geometric design: A practical guide [M].New York: Academic, 1990.
[24] J. M. Moulton. Spline surface interpolation for calculating three-dimensional ventricularstrains from MRI tissue tagging [J]. Physiology, 1996(270): 281~297.
[25] A. A. Amini. Energy-minimizing deformable grids for tracking tagged MR cardiac images
[A]. In: Computer in Cardiology [C]. Durham, North Carolina,USA, 1992：651~654.
[26] A. A. Amini, Y. Chen, W. R. Curwen, et al. Coupled B-snake grids and constrained thin-platesplines for analysis of 2D tissue deformations from tagged MRI [J]. IEEE Transactions MedicalImaging, 1998, 17(3): 344~356.
[27] J. Huang, D. Abendschein, G. V. Davila-Roman, et al. Spatial-Temporal tracking ofmyocardial deformations with a 4D B-spline model from tagged MRI [J]. IEEE Transactions MedicalImaging, 1999, 18(10): 957~972.
[28] C. Ozturk, R. E. McVeigh. Four-dimensional B-spline based motion analysis of taggedcardiac MR images: introduction and in vivo vallication [J]. Physics in Medicine and Biology, 2000,45(6):1683~1702.
[29] J. Declerck, J. Feldmar, N. Ayache. Definition of a 4D continuous planispherictransformation for the tracking and the analysis of LV motion [J]. Medical Image Analysis, 1998, 4(1):1~17.
[30] N. F. Osman, W. S. Kerwin, E.R. McVergh, et al. Cardiac motion tracking using CINEharmonic phase magnetic resonance imaging [J]. Magnetic Resonance in Medicine, 1999, 42(3):1048~1060.
[31] N. F. Osman, J. Prince. Visualizing mylcardial function using HARP MRI [J]. Physics inMedicine and Biology, 2000, 45(6):1665~1682.
[32] N. F. Osman, E. R. McVeigh, J. Prince. Imaging heart motion using harmonic phase MRI[J].IEEE Transactions Medical Imaging, 2000, 19(3): 186~202.
[33] N. F. Osman, J. L. Prince. Regenerating MR tagged images using harmonic phase (HARP)method [J]. IEEE Transactions on Biomedical Engineering, 2004, 51(8): 1428~1433.
[34] L. Pan, J. L. Prince, et al, Fast tracking of cardiac motion using 3D-HARP. IEEE transactionson Biomedical Engineering, 2005, 52(8):1424~1435.
[35] D. Gabor. Theory of communication [J]. Journal of the IEEE, 1946, 9(3): 429~457.
[36] W. T. Freeman, E. H. Adelson. The design and use of steerable filters[J]. IEEE Transactionson Pattern Analysis and Machine Intelligence, 1991, 13(9):891~906.
[37] C. L. Huang, Y. T. Chen. Motion estimation method using a 3D steerable filter[J]. Image andVision Computing, 1995, 13(1):.21~32.
[38] E. H. Adelson, J. R. Bergen. Spatiotemporal energy models for the perception of motion[J]. J.Opt. Soc. Am., 1995, 2(2): 284~299.
[39] E. P. Simoncelli, H. Farid. Steerable Wedge Filters for Local Orientation Analysis[J]. IEEETransactions on Image Processing, 1996, 5(9): 1377~1382.
[40] W. Yu, K. Daniilidis, G. Sommer. A new 3D orientation steerale filter[J]. In Proceedings ofDAGM-Symposium, 2000:203~212.
[41] J. M. Gryn. Three-dimensional nth derivative of gaussian separable steerable filters[J]. IEEEInternational Conference on Image Processing, 2005, 3:III–553-6.
[42] S. Xu, Y. Jia, X. Zhang. Coding facial expression with oriented steerable filters[J]. ICIP,2006: 2057~2060.
[43] N. Sahba, V. Tavakoli. Ultrasound thermal change detection based on steerable filters. 30thAnnual International IEEE EMBS Conference Vancouver, British Columbia, Canada, 2008:2984~2987.
[44] M. Muehlich, D. Friedrich, T. Aach. Design and implementation of Multi-Steerable MatchedFilters[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011.
[45] K. Krajsek, R. Mester. A unified theory for steerable and quadrature filters. Communicationsin Computer and Information Science, 2007, 4(10):.201~214.
[46] W. Zhou, H. Shu. Detection of Cerebral Vessels in MRA Using 3D Steerable Filters [J].Engineering in Medicine and Biology Society, 2005: 3249~3252.
[47] E. P. Simoncelli. The steerable pyramid: a flexible architecture for multi-scale derivativecomputation[C]. 2ndAnnual IEEE Conference on Image Processing, 1995,3: 444~447.
[48]顾宇灏，林青，胡波.一种基于方向可控金字塔结构的快速医学图像配准算法[J].复旦学报, 2010,49(2): 177~184.
[49]王蕊，尹忠科，龙奕.基于可控金字塔的轮廓波变换构造及其应用[J].铁道学报，2009,31(6): 53~57.
[50] X. Wang, D. D. Feng. Medical image registration via steerable pyramid[C]. Proceeding ofthe 29thAnnual International Conference of the IEEE EMBS, 2007: 6395~6398.
[51]顾立娟,邵命山,郝玉保.基于可控金字塔子带能量特征的文种识别方法[J].计算机应用与软件, 2011, 28(3):91~94.
[52]何青.利用方向可控金字塔进行基于互信息的图像配准[J].计算机应用, 2005, 25(12):2834~2836.
[53] B.K.P. Horn, B.G. Schunck. Determining optical flow [J]. Artificial Intelligence, 1981,17:185~204.
[54] M. Turner. Texture discrimination by gabor function [J]. Biological Cybernetics,1986,55:.71~82.
[55] Y. Zhu, T. Tan, Y. Wang. Biometric personal identification based on iris patterns [J].Proceedings of 15thInternational Conference on Pattern Recognition, 2000:801~804.
[56] C. Liu, H. Wechsler. Gabor feature based classification using the enhanced flasher linersdiscriminate model for face recognition[J]. IEEE Transactions on Image Processing, 2003, 11(2):.467~476.
[57] L. Hong, Y. Wan, A. Jain. Fingerprint image enhancement: algorithm and performanceevaluation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1998, 20(8): 777~789.
[58] A. Jain, K. N. Ratah, S. Lakshmanan. Object detection using gaborfilters[J]. PatternRecognition, 1997, 30(2): 295~309.
[59] D. N. Metaxas, L. Axel, Z. Qian, et al. A segmentation and Tracking System for 4D CardiacTagged MR Images[J]. IEEE EMBS Annual International Conference, 2006:1541~1544.
[60] L. Axel, S. Chung, T. Chen. Tagged MRI Analysis Using Gabor Filters[J]. IEEEInternational Symposium on Biomedical Imaging, 2007: 684~687.
[61] T. Chen, S. Chung, L. Axel. Automated Tag Tracking Using Gabor Filter Bank, Robust PointMatching, and Deformable Models[J]. FIMH, 2007: 22~31.
[62] T. Chen, X. Wang, D. Metaxas, et al. 3D Cardiac Motion Tracking Using Robust PointMatching and Models[J]. IEEE International Symposium on Biomedical Imaging, 2008:280~283.
[63] J. J. Koenderink, A. J. van Doorn. Representation of local geometry in the visual system[J].Biol. Cybern., 1987, 55: 367~375.
[64] H. Knutsson, R. Wilson, G. H. Granlund. Anisotropic non-stationary image estimation and itsapplications: Part 1-Restoration of noisy images[J]. IEEE Trans. Commun., 1983, 31(3):.388~397.