摘要
针对当前稀疏角度下有限角图像重建过程中,边界部分出现伪影,降低了图像重建质量的缺陷。文中提出了一种新的ART+TV算法,该方法是在原始TV算法的基础上进行改进。原始TV梯度下降算法求解目标函数最小值时,使用固定函数作为目标函数,文中对其进行更改,采用带参数的目标函数,并对TV重建后的结果进行自适应步长修正,加速图像收敛。与传统的ART+TV算法相比,文中算法在不改变重建速度的基础上,且在少量迭代次数下,能重建出质量更高的图像,抑制图像伪影。
For the current limited angle image reconstruction process sparse angle, the boundary part artifacts appear to reduce the image quality of the reconstructed defects. This paper proposes a new algorithm of ART ＋ TV, which is the basis of the original TV algorithms, to improve it. In solving the minimum objective function by the original TV gradient descent algorithm, the fixed function is used as the objective function. While in this paper, the text to change it, the objective function with parameters is employed, and the results of TV reconstruction are adaptive step corrected to accelerate the graphics convergence. With compare the proposed algorithm offers better quality of reconstructed images with a few iterations than the conventional ART ＋ TV algorithms at the same reconstruction speed and suppresses image artifacts
出处
《电子科技》
2016年第10期47-50,共4页
Electronic Science and Technology
