3D医学图像的脑血管提取与可视化
|
摘要
随着人们生活水平的提高,心脑血管疾病成为了影响人类健康最主要的病因之一,计算可视化技术及医学影像技术的发展,为心脑血管疾病的诊断输入了新的血液,脑血管的提取和可视化,使得医生从传统的诊断方法中脱离出来,不仅仅使医生获得了更丰富的医疗信息,更将人为的误差降低到最小。
针对图像中存在的噪声干扰,采用图像平滑算法对图像进行滤波,以此来消除图像中的噪声,平滑图像边缘；针对图像对比度差,采用了对比度增强算法增强了图像对比度；为了方便医生多视角观察切片组织,引入了几何变换算法,为血管提取及三维显示奠定基础。
血管轮廓的提取对于血管疾病的研究及血管中轴参数的确定是一个基础,对血管树的三维显示精度起着决定性的作用,纵观现有算法及其优缺点,综合采用了基于区域生长算法及基于Sobel算子的边缘检测算法,用于提取血管边缘轮廓。
血管中轴跟原图像具有相同拓扑性和相似的几何形状,因此能较好的反映血管的走向及空间结构,同时可以有效的佐证血管轮廓的提取的准确性。针对血管造影图像自身对比度差及背景复杂等缺陷,本文采用了多尺度Gabor变换算法提取血管中轴,在增强血管中轴的同时有效抑制图像噪声,通过调节方位带宽和频率得到最大响应图像,对响应图进行非极大值抑制,并通过阈值分割求出血管中轴。
利用三维重建技术,将提取后的血管图像进行三维还原,本文采用改进的移动立方体法(简称MC算法)进行血管的三维绘制,绘制结果表面平滑,没有空洞,达到了理想的效果。
With the improving of the life of the people, cardiovascular and cerebrovascular diseases become one main cause of disease which affect human health, the development of the computer graphics and visualization and the medical imaging join the new blood to the diagnosis of the cardiovascular and cerebrovascular diseases. The extraction and visualization of the Cerebral hemorrhage,which provide more medical information and less error, freed the doctor from the traditional diagnostic method.
In view of the noise in the data, smoothing filter is used to remove the noise and contrast enhancement is used to enhancement the contrast of the image, however, the geometric transformation algorithm are designed to provide the multi-angle of view for the medical worker, it set the stage for the extraction and the visualization of the blood.
The extraction of the blood edge,which play a critical part on the 3D reconstruction, lay the root for the research of the cerebrovascular and seeking for the parameter of the blood center shaft. In this thesis more algorithms were weighted, region growing algorithm and edge detection based on Sobel operator are used to detection the edge of the cerebrovascular.
The vessel axis can reflect well of the vessel direction and the space structure because of it having the same topology and geometrical, however, it can proved the efficiency of the vessel extraction.Multiple scales Gabor filter is used to extract the vessel axis owing to the complicated background and lower contrast of the angiography images, the vessel axis is enhanced while the noise is moved, we can get the maximal response diagram by adjusting the orientation bandwidth and the frequency bandwidth, after that, non-maximal rejection algorithm and threshold segmentation algorithm are adopted to detection the vessel axis.
3D reconstruct technology is used to restore the 3D structure of the extracted vessels. The improved moving cubes algorithm is (MC algorithm in short) adopted to plot the 3D vessels and the surface are smoothness without the hole as expected.
针对图像中存在的噪声干扰,采用图像平滑算法对图像进行滤波,以此来消除图像中的噪声,平滑图像边缘；针对图像对比度差,采用了对比度增强算法增强了图像对比度；为了方便医生多视角观察切片组织,引入了几何变换算法,为血管提取及三维显示奠定基础。
血管轮廓的提取对于血管疾病的研究及血管中轴参数的确定是一个基础,对血管树的三维显示精度起着决定性的作用,纵观现有算法及其优缺点,综合采用了基于区域生长算法及基于Sobel算子的边缘检测算法,用于提取血管边缘轮廓。
血管中轴跟原图像具有相同拓扑性和相似的几何形状,因此能较好的反映血管的走向及空间结构,同时可以有效的佐证血管轮廓的提取的准确性。针对血管造影图像自身对比度差及背景复杂等缺陷,本文采用了多尺度Gabor变换算法提取血管中轴,在增强血管中轴的同时有效抑制图像噪声,通过调节方位带宽和频率得到最大响应图像,对响应图进行非极大值抑制,并通过阈值分割求出血管中轴。
利用三维重建技术,将提取后的血管图像进行三维还原,本文采用改进的移动立方体法(简称MC算法)进行血管的三维绘制,绘制结果表面平滑,没有空洞,达到了理想的效果。
With the improving of the life of the people, cardiovascular and cerebrovascular diseases become one main cause of disease which affect human health, the development of the computer graphics and visualization and the medical imaging join the new blood to the diagnosis of the cardiovascular and cerebrovascular diseases. The extraction and visualization of the Cerebral hemorrhage,which provide more medical information and less error, freed the doctor from the traditional diagnostic method.
In view of the noise in the data, smoothing filter is used to remove the noise and contrast enhancement is used to enhancement the contrast of the image, however, the geometric transformation algorithm are designed to provide the multi-angle of view for the medical worker, it set the stage for the extraction and the visualization of the blood.
The extraction of the blood edge,which play a critical part on the 3D reconstruction, lay the root for the research of the cerebrovascular and seeking for the parameter of the blood center shaft. In this thesis more algorithms were weighted, region growing algorithm and edge detection based on Sobel operator are used to detection the edge of the cerebrovascular.
The vessel axis can reflect well of the vessel direction and the space structure because of it having the same topology and geometrical, however, it can proved the efficiency of the vessel extraction.Multiple scales Gabor filter is used to extract the vessel axis owing to the complicated background and lower contrast of the angiography images, the vessel axis is enhanced while the noise is moved, we can get the maximal response diagram by adjusting the orientation bandwidth and the frequency bandwidth, after that, non-maximal rejection algorithm and threshold segmentation algorithm are adopted to detection the vessel axis.
3D reconstruct technology is used to restore the 3D structure of the extracted vessels. The improved moving cubes algorithm is (MC algorithm in short) adopted to plot the 3D vessels and the surface are smoothness without the hole as expected.
引文
[1]田捷,包尚联等.医学影像处理技术与分析[M].北京：电子工业出版社,2003.
[2]李辉.计算机辅助外科得现状和前景[J].中国外科志.2004,42(11)：689-699.
[3]Ommer O Dietz A, Westermann R, An interactive visualization and navigation Tool for medical volume data[J]. Computers and Graphics.1999,23:233-244
[2]李辉.计算机辅助外科的现状和前景[J].中国外科志.2004,42(11)：689-699.
[3]Ommer O Dietz A, Westermann R, An interactive visualization and navigation tool for medical volume data[J]. Computers and Graphics.1999,23:233-244.
[4]Sarti A, Lamberti C, Erbacci G. Volume redering for 3D echocardiography visualization[J]. Proceedings of Computers in Gardiology.1993,9:209-212.
[5]郑国炎,江贵平等.一种计算机辅助手术方案仿真系统[J].中国生物医学工程学报.1998,17(2)：151-159.
[6]ACR-NEMA Commitee.Digital Imaging and Communication in medicine[J].2009.
[7]Eberly D, Gardner R, Morse R. Ridges for image analysis[J].Journal of Mathematical Imaging and Vision.2009,4:353-373.
[8]Bullitt E,Aylward S.Analysis of time-varying images using 3D vascular models[J]. Proc.of Applied Imagery Pattern Recogonition Workshop, Washington, DC,USA. 2001:9-14.
[9]Schmitt H, Grass M, Rasche V. An x-ray-based method for the determination of the contrast agent propagation in 3D vessel structures[J]. IEEE Transactions on Medical Imaging.2002,21(3):251-262.
[10]冈萨雷斯.数字图像处理[M].北京：电子工业出版社.
[11]Do Carmo M. Differential geometry of curves and surfaces[J]. NewJersy:Prentice-hall EngleWood Cliffs.
[12]Chwialkowski M,Ibrahim Y,Hong F,et.A method for fully automated quantitative analysis of arterial using flow-sensitized MR images[J]. Computeriezed Medical Imaging and Graphics.1996,20(5):365-378.
[13]刘潇潇,曹治国,李抱朴等.基于多尺度Gabor滤波的造影血管中轴线的自动提取[J].中国图象图形学报.2005,12(10)：1542-1546.
[14]Sarwal A, Dhawan A.3D reconstruction of coronary arteries[J].Proc.of IEEE Conference Engineering in Medicine and Biolagy Society,1994,1:504-505.
[15]Gullberg G,Zeng G. A cone-beam filtered backpropagation reconstruction algorithm for cardiac single photon emission computed tomography[J].IEEE Transactions on Medical Imaging.1992,Mi-11:91-101.
[16]Obrien J,Ezquerra N.Automated segmentation of coronary vessels in angiographic image sequences utilizing temporal,spatial structural constraints[J].Proc.of SPIE Conference Visualization in Biomedical Computing.1994,2359(25):25-37.
[17]Ripley B. Pattern recognition and neural networks[M]. British:Cambridge University press,1996.
[18]Haykin S. Neural networks:a comprehensive foundation[M]. USA:Printice Hall,2002.
[19]Shiffman S,Rubin G. Semiautomated editing of computed tomography sections for visualization of vascular[J]. Proc of SPIE Medical Imaging Conference.1996,2507:140-151.
[20]Kass M,Witkin A.Snakes:active contour models[J].International Joumal of Computer Vision.1988,1(4):321-331.
[21]Smete C,Verbecck G.A knowledge-based system for the delincation of blood vessels on subtraction angiograms[J]. Pattern Recognition Letters.2001,8(2):113-121.
[22]Liu S, Sun E.A automatic detection of the coronary arterial contours with sub-branches from an x-ray angiogram[J]. Proc of IEEE Computers in cardiology,London,UK.2001,8:575-.578.
[23]Park S. Adaptive tracking algorithm based on direction field using ML estimation in angiogram[J]. Proc. of IEEE Conference on speech and Image Technologies for Computing and Telecommunications, Brisbane, Austrilia.1993,9:575-578.
[24]V. Caselles, F. Cate. A geometric model for active contours in image processing[J]. Nm.processing.Nm.1933,66:1-32.
[25]R. Malladi, J. A. Sethian. Shape modeling with front propagation:A level set approach PAMI.1995,17:158-175.
[26]彭群生,金小刚等.计算机图形学应用基础[M].北京：科学出版社.
[27]廖秀秀,梁礼健.医学图像三维重建技术.中国医学装备[J].2009,2(5)：21-23.
[28]金天弘,刘振宅.医学图像三维重建的研究.医疗卫生装备[J].2008,2(29)34-36.
[29]李艳西.血管造影图像的中轴提取技术研究[D].硕士论文.
[30]胡学龙,许开于.数字图像处理[M].北京：电子工业出版社.
[31]张娜.数字图像增强的方法.大众科技,2006,8：27-28.
[32]孙兆林.MATLAB 6.x图像处理[M].北京：清华大学出版社.
[33]徐丹红,王保华,张勇,沈海东等.基于区域生长分割的三维心腔重建[J].中国医疗器械杂志,2007,31(1)：17-21.
[34]章毓晋.图像分割[M].北京：清华大学出版社,2000.
[35]吴敏金.图像形态学[M].上海：上海科技技术文献,1991.
[36]田捷等.实用图像处理技术[M].北京：电子工业出版社,1994.
[37]K.Palagyi, A. Kuba, Directional 3D Thinning using 8 Subiterations[J]. Proc.DGCT99, LNCS1568,1999:325-326.
[38]Rosenfeld, Pfaltz J L. Sequential operations in digital picture processing[J]. Journal of ACM,1966,13(4):471-494.
[39]李颖超,刘越,王涌天.基于多尺度Hessian矩阵和Gabor滤波的造影图像冠状中心线提取[J].中国医学影像技术,2007,23(1)：133-136.
[40]Khaled Hammouda Prof. Ed Jernigan. Texture Segmentation Using Gabor Filters[J].Department of Systems Design Engineering.2000,12.
[41]John Canny. A Computational Approach to Edge Detection[J]. IEEE Transactions on Pattern analysis and machine intelligence.1986:679-698.
[42]Lorensen W. E, Cline H. E. Marching cube:a high resolution 3D surface construction algorithm. Computer Graphics,1987,21(4),163-169.
[43]Durst M J. Letters:additional Reference to "Marching Cubes"[J]. Computer Graphics,1988,22(2):72-73.
[44]E. Keppel. Approximating complex surface by triangulation of contour lines [J].IBM Joural of research and development,1975,19(1):2-11.
[45]张尤赛,陈福民.三维医学图像的体绘制技术综述.计算机工程与应用.2002.8：18-20.
[46]Levoy M. Display of Surfaces from Volume Data[J]. IEEE Computer Graphics and Application.1988,8(3):29-37.
[47]Westover L. Footprint evaluation for volume rendering [J]. Computer Graphics.1990,24(4):367-376.
[48]Lacroute, M. Levoy. Fast Volume Rendering using a shear-warp factorization of the viewing transformation[J]. Computer Graphics 1994:451-457.
[49]唐泽圣.三维数据场可视化[M].北京：清华大学出版社.1999,12.
[50]Bcavral, N. Accelerated volume rendering and tomographic reconstruction using texture mapping hardware. Symposium on Visualization.1994:91-98.
[51]李华,蒙陪生,王乘.医学图像重建MC算法三角片的合并与实现.计算机应用[J].2003,23(6)：104-106.
[52]管伟光.体视化技术及其应用[M].北京：电子工业出版社.1998.
[53]Nielson G M, Hamann B. The Asumptotic Decider:Resolving the Ambiguity in Marching Cubes[J]. IEEE Proceeding of Visualization.1991:83-91
[2]李辉.计算机辅助外科得现状和前景[J].中国外科志.2004,42(11)：689-699.
[3]Ommer O Dietz A, Westermann R, An interactive visualization and navigation Tool for medical volume data[J]. Computers and Graphics.1999,23:233-244
[2]李辉.计算机辅助外科的现状和前景[J].中国外科志.2004,42(11)：689-699.
[3]Ommer O Dietz A, Westermann R, An interactive visualization and navigation tool for medical volume data[J]. Computers and Graphics.1999,23:233-244.
[4]Sarti A, Lamberti C, Erbacci G. Volume redering for 3D echocardiography visualization[J]. Proceedings of Computers in Gardiology.1993,9:209-212.
[5]郑国炎,江贵平等.一种计算机辅助手术方案仿真系统[J].中国生物医学工程学报.1998,17(2)：151-159.
[6]ACR-NEMA Commitee.Digital Imaging and Communication in medicine[J].2009.
[7]Eberly D, Gardner R, Morse R. Ridges for image analysis[J].Journal of Mathematical Imaging and Vision.2009,4:353-373.
[8]Bullitt E,Aylward S.Analysis of time-varying images using 3D vascular models[J]. Proc.of Applied Imagery Pattern Recogonition Workshop, Washington, DC,USA. 2001:9-14.
[9]Schmitt H, Grass M, Rasche V. An x-ray-based method for the determination of the contrast agent propagation in 3D vessel structures[J]. IEEE Transactions on Medical Imaging.2002,21(3):251-262.
[10]冈萨雷斯.数字图像处理[M].北京：电子工业出版社.
[11]Do Carmo M. Differential geometry of curves and surfaces[J]. NewJersy:Prentice-hall EngleWood Cliffs.
[12]Chwialkowski M,Ibrahim Y,Hong F,et.A method for fully automated quantitative analysis of arterial using flow-sensitized MR images[J]. Computeriezed Medical Imaging and Graphics.1996,20(5):365-378.
[13]刘潇潇,曹治国,李抱朴等.基于多尺度Gabor滤波的造影血管中轴线的自动提取[J].中国图象图形学报.2005,12(10)：1542-1546.
[14]Sarwal A, Dhawan A.3D reconstruction of coronary arteries[J].Proc.of IEEE Conference Engineering in Medicine and Biolagy Society,1994,1:504-505.
[15]Gullberg G,Zeng G. A cone-beam filtered backpropagation reconstruction algorithm for cardiac single photon emission computed tomography[J].IEEE Transactions on Medical Imaging.1992,Mi-11:91-101.
[16]Obrien J,Ezquerra N.Automated segmentation of coronary vessels in angiographic image sequences utilizing temporal,spatial structural constraints[J].Proc.of SPIE Conference Visualization in Biomedical Computing.1994,2359(25):25-37.
[17]Ripley B. Pattern recognition and neural networks[M]. British:Cambridge University press,1996.
[18]Haykin S. Neural networks:a comprehensive foundation[M]. USA:Printice Hall,2002.
[19]Shiffman S,Rubin G. Semiautomated editing of computed tomography sections for visualization of vascular[J]. Proc of SPIE Medical Imaging Conference.1996,2507:140-151.
[20]Kass M,Witkin A.Snakes:active contour models[J].International Joumal of Computer Vision.1988,1(4):321-331.
[21]Smete C,Verbecck G.A knowledge-based system for the delincation of blood vessels on subtraction angiograms[J]. Pattern Recognition Letters.2001,8(2):113-121.
[22]Liu S, Sun E.A automatic detection of the coronary arterial contours with sub-branches from an x-ray angiogram[J]. Proc of IEEE Computers in cardiology,London,UK.2001,8:575-.578.
[23]Park S. Adaptive tracking algorithm based on direction field using ML estimation in angiogram[J]. Proc. of IEEE Conference on speech and Image Technologies for Computing and Telecommunications, Brisbane, Austrilia.1993,9:575-578.
[24]V. Caselles, F. Cate. A geometric model for active contours in image processing[J]. Nm.processing.Nm.1933,66:1-32.
[25]R. Malladi, J. A. Sethian. Shape modeling with front propagation:A level set approach PAMI.1995,17:158-175.
[26]彭群生,金小刚等.计算机图形学应用基础[M].北京：科学出版社.
[27]廖秀秀,梁礼健.医学图像三维重建技术.中国医学装备[J].2009,2(5)：21-23.
[28]金天弘,刘振宅.医学图像三维重建的研究.医疗卫生装备[J].2008,2(29)34-36.
[29]李艳西.血管造影图像的中轴提取技术研究[D].硕士论文.
[30]胡学龙,许开于.数字图像处理[M].北京：电子工业出版社.
[31]张娜.数字图像增强的方法.大众科技,2006,8：27-28.
[32]孙兆林.MATLAB 6.x图像处理[M].北京：清华大学出版社.
[33]徐丹红,王保华,张勇,沈海东等.基于区域生长分割的三维心腔重建[J].中国医疗器械杂志,2007,31(1)：17-21.
[34]章毓晋.图像分割[M].北京：清华大学出版社,2000.
[35]吴敏金.图像形态学[M].上海：上海科技技术文献,1991.
[36]田捷等.实用图像处理技术[M].北京：电子工业出版社,1994.
[37]K.Palagyi, A. Kuba, Directional 3D Thinning using 8 Subiterations[J]. Proc.DGCT99, LNCS1568,1999:325-326.
[38]Rosenfeld, Pfaltz J L. Sequential operations in digital picture processing[J]. Journal of ACM,1966,13(4):471-494.
[39]李颖超,刘越,王涌天.基于多尺度Hessian矩阵和Gabor滤波的造影图像冠状中心线提取[J].中国医学影像技术,2007,23(1)：133-136.
[40]Khaled Hammouda Prof. Ed Jernigan. Texture Segmentation Using Gabor Filters[J].Department of Systems Design Engineering.2000,12.
[41]John Canny. A Computational Approach to Edge Detection[J]. IEEE Transactions on Pattern analysis and machine intelligence.1986:679-698.
[42]Lorensen W. E, Cline H. E. Marching cube:a high resolution 3D surface construction algorithm. Computer Graphics,1987,21(4),163-169.
[43]Durst M J. Letters:additional Reference to "Marching Cubes"[J]. Computer Graphics,1988,22(2):72-73.
[44]E. Keppel. Approximating complex surface by triangulation of contour lines [J].IBM Joural of research and development,1975,19(1):2-11.
[45]张尤赛,陈福民.三维医学图像的体绘制技术综述.计算机工程与应用.2002.8：18-20.
[46]Levoy M. Display of Surfaces from Volume Data[J]. IEEE Computer Graphics and Application.1988,8(3):29-37.
[47]Westover L. Footprint evaluation for volume rendering [J]. Computer Graphics.1990,24(4):367-376.
[48]Lacroute, M. Levoy. Fast Volume Rendering using a shear-warp factorization of the viewing transformation[J]. Computer Graphics 1994:451-457.
[49]唐泽圣.三维数据场可视化[M].北京：清华大学出版社.1999,12.
[50]Bcavral, N. Accelerated volume rendering and tomographic reconstruction using texture mapping hardware. Symposium on Visualization.1994:91-98.
[51]李华,蒙陪生,王乘.医学图像重建MC算法三角片的合并与实现.计算机应用[J].2003,23(6)：104-106.
[52]管伟光.体视化技术及其应用[M].北京：电子工业出版社.1998.
[53]Nielson G M, Hamann B. The Asumptotic Decider:Resolving the Ambiguity in Marching Cubes[J]. IEEE Proceeding of Visualization.1991:83-91