摘要
本文采用连续型Hopfield神经冈络(HNN)模型应用于有限角CT图像重建,该重建算法将两种优化准则转化为模型的能量系数求取图像重建的最优解,通过数值模拟,考察了该重建算法和抗噪性能,均得到了良好的结果。
A new algorithm of the image reconstruction on limited view CT using continuous HNN was presented. Two kinds of criteria were changed into the energy function in HNN to find the best solution. The numerical simulation was used to study the effect of image reconstruction and stability against noise. The results show that this algorithm has good accuracy.
出处
《光电子技术与信息》
CAS
2005年第2期74-77,共4页
Optoelectronic Technology & Information
关键词
有限角CT
神经网络
图像重建
limited view CT
neural network
image reconstruction
