基于各向异性扩散和纠正探测值的PET图像重建  

A PET Imaging Reconstruction Based on Anisotropic Diffusion and Correction of Detected Sinogram Data

作　　者：王红红[1] 桂志国[1] 

机构地区：[1]中北大学电子测试技术国家重点实验室,山西太原030051

出　　处：《测试技术学报》2011年第5期461-465,共5页Journal of Test and Measurement Technology

基　　金：山西省自然科学基金资助项目(2009011020-2)

摘　　要：针对如何抑制正电子发射成像(positron emission tomography,PET)中噪声的问题,提出了一种改进的最大化后验估计(maximum a posteriori,MAP)算法,该算法将耦合反馈(coupled feedback,CF)模型和各向异性扩散(anisotropic diffusion,AD)模型结合起来应用在MAP算法中.在重建图像过程中,用AD模型对上次迭代重建的图像进行去噪,根据去噪后的图像用CF模型得到纠正信息,再用此纠正信息更新有噪声的探测数据.模拟实验的结果证明,对于PET重建,该算法降低了重建图像的噪声,同时有效地保护了重建图像的细节与边缘.Aiming at the problem of suppressing noises in reconstructed images of positron emission tomography（PET）, we put forward an improved algorithm of maximum a poster（MAP） by adopting anisotropic diffu- sion（AD） model combining with coupled feedback（CF） model. In the process of reconstructing image, the AD model was firstly used to remove the noises in the image reconstructed from last iteration, and then the CF model was used to obtain the correction information using the denoised image to update the noisy detected data The simulated experiment results indicate that this algorithm not only removes noise but also protects image details and edges efficiently for PET reconstruction.

关 键 词：正电子发射成像 最大化后验估计 耦合反馈迭代模型 探测数据 各向异性扩散 

分 类 号：TN919.8[电子电信—通信与信息系统]