图像配准在高频超声皮肤三维成像中的应用研究
|
摘要
目的:将图像配准技术应用在高频超声皮肤三维成像系统中,消除由于三维探头的振动、操作者和受试者的不自主运动等给三维数据带来的偏差,提高高频超声三维成像的精度。方法:以C++语言为基础,结合图像处理工具包ITK(insight segmentation and registration toolkit),在保持配准框架其他组成部分不变的情况下,仅改变变换函数对高频超声皮肤图像进行配准实验。通过可视化工具包VTK(visualization toolkit)观察三维数据配准前后三维重建图像的表面情况及其内部切面图,并选用相关系数、结构相似性指数、余弦相似度以及配准时间尝试对配准效果进行定量评价。结果:平移变换、刚性变换、二维相似变换以及仿射变换的配准对三维重建图像质量有明显改善,B样条可变形变换配准效果不理想。5种变换函数中,刚性变换配准效果最好;平移变换配准和刚性变换配准效果接近,且所需配准时间最少。结论:图像配准能显著提高高频超声三维重建图像的精度,在兼顾配准时间和精度的情况下,平移变换是高频超声皮肤三维成像中较为理想的图像配准方法。
Objective To apply image registration technology to high-frequency ultrasound skin three-dimensional imaging system so as to eliminate deviations from three-dimensional data due to vibrations of the three-dimensional probe, involuntary movements of the operator and the subject and improve the accuracy of high-frequency ultrasound three-dimensional imaging.Methods The registration experiment was based on C ++ language, combined with the image processing toolkit ITK, only changing the transform function to register the high frequency ultrasonic skin image while keeping the other parts of the registration frame unchanged. VTK was used to observe the surface and internal section of three-dimensional reconstructed image before and after the registration of three-dimensional data. The three similarity measures of the correlation coefficient,the structural similarity index, cosine similarity and the registration time were used to quantitatively evaluate the registration effect. Results The registration of translation transformation, rigid transformation, two-dimensional similar transformation and affine transformation improved the image quality of three-dimensional reconstruction, and the B-spline deformable transformation registration effect was not ideal. Among the five kinds of transformation functions, the registration effect of rigid transformation was the best, the registration effect of translation transformation was close to that of rigid transformation, and the registration time was the least. Conclusion Image registration can significantly improve the accuracy of three-dimensional reconstruction of high-frequency ultrasound images. In the case of both registration time and accuracy, translational transformation is an ideal image registration method for high-frequency ultrasound skin three-dimensional imaging. [Chinese Medical Equipment Journal,2019,40(7):6-10]
Objective To apply image registration technology to high-frequency ultrasound skin three-dimensional imaging system so as to eliminate deviations from three-dimensional data due to vibrations of the three-dimensional probe, involuntary movements of the operator and the subject and improve the accuracy of high-frequency ultrasound three-dimensional imaging.Methods The registration experiment was based on C ++ language, combined with the image processing toolkit ITK, only changing the transform function to register the high frequency ultrasonic skin image while keeping the other parts of the registration frame unchanged. VTK was used to observe the surface and internal section of three-dimensional reconstructed image before and after the registration of three-dimensional data. The three similarity measures of the correlation coefficient,the structural similarity index, cosine similarity and the registration time were used to quantitatively evaluate the registration effect. Results The registration of translation transformation, rigid transformation, two-dimensional similar transformation and affine transformation improved the image quality of three-dimensional reconstruction, and the B-spline deformable transformation registration effect was not ideal. Among the five kinds of transformation functions, the registration effect of rigid transformation was the best, the registration effect of translation transformation was close to that of rigid transformation, and the registration time was the least. Conclusion Image registration can significantly improve the accuracy of three-dimensional reconstruction of high-frequency ultrasound images. In the case of both registration time and accuracy, translational transformation is an ideal image registration method for high-frequency ultrasound skin three-dimensional imaging. [Chinese Medical Equipment Journal,2019,40(7):6-10]
引文
[1]王艺,韩若凌,高洁宁.三维超声技术进展及其在浅表器官诊断中的应用[J].中华医学超声杂志:电子版,2015,12(3):184-187.
[2] GEBHARDR E,EUBANKST N,MEEKS R,et al. Threedimensional ultrasound imaging[J]. Curr Opin Anaesthesiol,2015,28:583-587.
[3] MANI V R S,SELVARAJ A. Survey of medical image registration[J]. IJBET,2013,1(2):8-25.
[4] ZHEN L. Feature-based image registration using the shape context[J]. Int J Remote Sens,2010,31(8):2 169-2 177.
[5]韦春荣,周永健,张灿龙.基于频域时域相结合的医学图像配准算法[J].中国组织工程研究,2009,13(13):2 582-2 584.
[6] SCHNEIDER R J,PERRIN D P,VASILYEV N V,et al. Realtime image-based rigid registration of three-dimensional ultrasound[J]. Med Image Anal,2012,16(2):402-414.
[7] AVANTS B B,TUSTISON N J,MICHAEL S,et al. The Insight ToolKit image registration framework[J]. Front Neuroinform,2014,8:44-57.
[8]赵菲,廖琪梅,马昂昂,等.基于ITK的多模态医学图像非刚性配准研究[J].医疗卫生装备,2011,32(11):5-8.
[9] LIU Y,CHENG H D,HUANG J H,et al. An effective nonrigid registration approach for ultrasound image based on“demons”algorithm[J]. J Digit Imaging,2013,26(3):521-529.
[10]吕晓琪,张宝华,杨立东,等.医学图像配准技术与应用[M].北京:科学出版社,2015:56-58.
[11]杭利华,蒋佩钊,邓冲.基于仿射变换与B样条自由形变的医学图像配准[J].兰州交通大学学报,2013,32(3):44-48,65.
[12]张倩,朱虹,郑丽敏,等.几种相似性测度在小鼠序列图片中的应用对比[J].计算机应用,2011,31(S1):60-63.
[13] PLUIM J P,MAINTZ J B,VIERGEVER M A. Mutual-information-based registration of medical images:a survey[J]. IEEE Trans Med Imag,2003,22(8):986-1 004.
[14]王文赛,王韬,周榴,等.基于FPGA的高频超声皮肤三维扫描控制及数据采集系统设计[J].医疗卫生装备,2018,39(11):5-9.
[2] GEBHARDR E,EUBANKST N,MEEKS R,et al. Threedimensional ultrasound imaging[J]. Curr Opin Anaesthesiol,2015,28:583-587.
[3] MANI V R S,SELVARAJ A. Survey of medical image registration[J]. IJBET,2013,1(2):8-25.
[4] ZHEN L. Feature-based image registration using the shape context[J]. Int J Remote Sens,2010,31(8):2 169-2 177.
[5]韦春荣,周永健,张灿龙.基于频域时域相结合的医学图像配准算法[J].中国组织工程研究,2009,13(13):2 582-2 584.
[6] SCHNEIDER R J,PERRIN D P,VASILYEV N V,et al. Realtime image-based rigid registration of three-dimensional ultrasound[J]. Med Image Anal,2012,16(2):402-414.
[7] AVANTS B B,TUSTISON N J,MICHAEL S,et al. The Insight ToolKit image registration framework[J]. Front Neuroinform,2014,8:44-57.
[8]赵菲,廖琪梅,马昂昂,等.基于ITK的多模态医学图像非刚性配准研究[J].医疗卫生装备,2011,32(11):5-8.
[9] LIU Y,CHENG H D,HUANG J H,et al. An effective nonrigid registration approach for ultrasound image based on“demons”algorithm[J]. J Digit Imaging,2013,26(3):521-529.
[10]吕晓琪,张宝华,杨立东,等.医学图像配准技术与应用[M].北京:科学出版社,2015:56-58.
[11]杭利华,蒋佩钊,邓冲.基于仿射变换与B样条自由形变的医学图像配准[J].兰州交通大学学报,2013,32(3):44-48,65.
[12]张倩,朱虹,郑丽敏,等.几种相似性测度在小鼠序列图片中的应用对比[J].计算机应用,2011,31(S1):60-63.
[13] PLUIM J P,MAINTZ J B,VIERGEVER M A. Mutual-information-based registration of medical images:a survey[J]. IEEE Trans Med Imag,2003,22(8):986-1 004.
[14]王文赛,王韬,周榴,等.基于FPGA的高频超声皮肤三维扫描控制及数据采集系统设计[J].医疗卫生装备,2018,39(11):5-9.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |