自适应正则MAP的CT图像重建方法研究

摘要

This method which is an adaptive regularized MAP of CT image reconstruction can adaptively choose regulariza-tion parameters with making full use of the results of every iteration to update regularization parameters and gets the final reconstruction image through multiple iterations.It simulates a slice of Shepp-Logan model to verify the feasibility of the al-gorithm.The results show that compared with the traditional methods, ML-EM and FBP,the method is more applicable, can better maintain the image edge information,has a higher signal to noise ratio images and improves the image quality.%提出了一种基于自适应正则最大后验概率估计(MAP)的计算机断层(CT)图像重建方法,该方法可以自适应地选择正则化参数,并充分利用每一次迭代的重建结果的信息,不断对其进行更新,通过多次迭代得到最终的重建图像.通过对头部模型的一个切片进行仿真实验,验证了该方法的可行性.用实际实验数据,与传统的方法、最大似然期望(ML-EM)算法和滤波反投影(FBP)算法相比较,表明该方法适用性较强,能较好地保持图像的边缘信息,而且图像的信噪比较高.