| 层析成像的模糊-神经网络重建方法 |
| |
| 引用本文： | 李海洋,张建中,徐初伟. 层析成像的模糊-神经网络重建方法[J]. 厦门大学学报(自然科学版), 2005, 44(3): 330-332 |
| |
| 作者姓名： | 李海洋 张建中 徐初伟 |
| |
| 作者单位： | 厦门大学信息科学与技术学院,福建,厦门,361005 |
| |
| 基金项目： | 福建省自然科学基金(D0310001)，厦门大学科研基金 |
| |
| 摘 要： | 考虑投影数据不足时，射线层析成像的离散重建算法难以取得较好效果的情况，利用模糊理论建立了表示物体衰减系数的连续函数，用神经网络学习算法确定该连续函数中的待定参数．提出了一种适于投影数据或投影方向不够时的CT图像重建算法．计算机仿真结果显示了该方法在投影数据严重缺少时的良好效果。
|
| 关 键 词： | 层析成像 重建方法 神经网络 投影数据 图像重建算法 网络学习算法 连续函数 衰减系数 理论建立 仿真结果 投影方向 待定参数 计算机 |
| 文章编号： | 0438-0479(2005)03-0330-03 |
| 修稿时间： | 2004-04-26 |
A Neuro-fuzzy Reconstruction Algorithm for Tomography |
| |
| Li Hai-Yang,ZHANG Jian-zhong,XU Chu-wei. A Neuro-fuzzy Reconstruction Algorithm for Tomography[J]. Journal of Xiamen University(Natural Science), 2005, 44(3): 330-332 |
| |
| Authors: | Li Hai-Yang ZHANG Jian-zhong XU Chu-wei |
| |
| Affiliation: | LI Hai-yang,ZHANG Jian-zhong~*,XU Chu-wei |
| |
| Abstract: | The discrete reconstruction algorithm for ray CT divides the object into many non-overlapping pixels.In order to obtain the good result of reconstruction,each pixel must be enough small,and be covered with a number of rays.But when the projection data are scarce,two requirements can't be met at the same time.In this paper,we have proposed a novel CT image reconstruction algorithm that still holds true when the projection data or directions are deficient.According to the proposed algorithm,we make use of the fuzzy theory to set up the continuous function of attenuation rate of the detected object,and determine the undetermined parameters of the continuous function by means of the learning algorithm of the neural network.Computer simulation results show the good effect of the algorithm when the projection data are seriously scarce. |
| |
| Keywords: | CT fuzzy logic neural network inversion |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
