摘要
光声成像结合了光学成像和声成像的优点,是一种具有高空间分辨率、高对比度的无损成像技术,成为当前生物医学成像的研究热点之一。重建光声图像是一个典型的逆问题,具有严重的病态性。针对光声成像的病态性和较大的系统矩阵会导致重建速度慢的问题,提出了一种基于Lanczos双对角化的快速指数滤波重建方法,并通过数值仿真证实了该方法的有效性。仿真结果表明,所提方法在保证重建图像高质量的同时极大地提高了重建速度,其重建时间是指数滤波和后投影方法的1/67~1/47。
Photoacoustic imaging, which combines the advantages of optical imaging and acoustic imaging, is a non- invasive imaging technique with high spatial resolution and high contrast ratio. Therefore, photoacoustic imaging is one of the research hot spots in biomedical imaging. However, the reconstruction of photoacoustic image is a typically ill-posed inverse problem. Aiming at the ill-posedness of photoacoustic imaging and the slow velocity of the reconstruction owing to the large size of the system matrix, we present a fast exponential-filtering reconstruction method based on Lanczos double diagonalization and the effectiveness of the proposed method is proved by numerical simulation. The simulation results show that the proposed method can increase the reconstruction velocity greatly, while preserving the high quality of reconstructed images. The reconstruction time is 1/67-1/47 of those using exponential-filtering method and back projection method.
作者
张娜
冯金超
李哲
贾克斌
Zhang Na1, Feng Jinchao1,2, Li Zhe1,2, Jia Kebin 1,2,3(1Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China Z Beij ing Laboratory of Advanced Information Networks, Beij ing 100124, China ; 3 Beijing Advanced Innovation Center for Future Internet Technology, Beijing University of Technology, Beijing 100124, Chin)
出处
《中国激光》
EI
CAS
CSCD
北大核心
2018年第3期165-173,共9页
Chinese Journal of Lasers
基金
国家自然科学基金(81370038)
北京市科技新星计划(Z141101001814107)
中国博士后基金(2015T80030
2016M600884)
先进信息网络北京实验室(040000546617002)
关键词
医用光学
图像处理
光声成像
正则化
指数滤波
Lanczos双对角化
medical optics
image processing
photoacoustic imaging
regularization
exponential filtering
Lanczosdouble diagonalization