MART算法快速高质量图像重建研究  

Fast and High Quality Image Reconstruction for Multiplicative ART

作　　者：梅创社[1] 张顺利[2] 

机构地区：[1]陕西工业职业技术学院,咸阳712000 [2]西北工业大学现代设计与集成制造技术教育部重点实验室,西安710072

出　　处：《科学技术与工程》2012年第24期6054-6058,共5页Science Technology and Engineering

基　　金：陕西省自然科学基金(2009JQ8017);陕西省教育厅专项基金(11JK1030)资助

摘　　要：ART(Algebraic Reconstruction Technique)算法是一种典型的迭代图像重建算法,适合于不完全投影数据图像重建。为了提高乘型ART(Multiplicative ART,MART)算法的重建质量,提出了一种基于亚像素的图像重建方法。首先将原始图像中的每个像素等分解为四个亚像素,然后提出一种高效的射线与像素的求交算法来计算权因子和亚像素索引。利用MART算法重建得到高分辨率重建图像。最后通过合并亚像素图像得到原始分辨率的高质量重建图像。实验结果表明提出的方法非常有效,与传统方法相比取得了3倍以上的重建加速比,图像重建质量显著提高。The algebraic reconstruction technique （ART） is a classical iterative image reconstruction algo- rithm, which is suit for image reconstruction from incomplete projection data. In order to improve the image recon-struction quality of MART （Multiplicative ART） , an image reconstruction method based on sub-pixel is presented. Firstly, each pixel of original image is equally decomposed into four sub-pixels. Then a fast intersection algorithm of ray-pixel is presented to calculate weight coefficients and index of sub-pixel. The reconstructed high resolution image is obtained using MART, from which the high quality reconstruction image with original resolution by merging sub-pixels can be obtained. Experimental result shows that the proposed method is very effective, compared with traditional method, can be got a reconstruction speedup over 3 times and improve the reconstructed image quality obviously.

关 键 词：CT MART 图像重建 亚像素 

分 类 号：TP391.74[自动化与计算机技术—计算机应用技术]