最大后验概率重建算法在发射CT中的应用

详细信息

Maximum a posteriori Reconstruction Algorithms in Emission CT

Institute of Nuclear and New energy Technology, Tsinghua University, Beijing 100084

摘要: 最大似然期望最大(MaximumLikelihood-ExpectationMaximization,MLEM)重建算法已在发射CT的二维重建中得到了广泛的研究和应用,但MLEM有两点不足:1)一定迭代次数以后噪声随着迭代增多而增多。2)重建需要太长时间。而最大后验概率重建算法(maximumaposteriori,MAP)根据图像的先验知识分布,则可以有效的抑制噪声,后来逐渐发展起来的多种加速迭代方法结合MAP重建算法可以缩短重建时间。本文对近年来MAP重建算法的发展做了一个回顾,并对各类算法了做了比较和讨论。
- 发射CT /
- 先验知识分布 /
- MAP重建算法
Abstract: MLEM (Maximum Likelihood-Expectation Maximization) has been widely used in Emission CT, however, in practice two draw backs are encountered: 1) due to the ill-conditioned nature of the reconstruction problem, reconstructions tent to take on an increasing non-smooth quality as the ML solution is approached; 2) convergence can be increasingly slow as the ML solution is approached. Maximum a posteriori (MAP) reconstruction has been shown to have significant advantages over MLEM in terms of noise reduction and the convergence speed. A review of development of MAP algorithms is presented in this paper.
- Emission CT /
- prior distribution /
- MAP reconstruction algorithms