摘要
目的研究FitzHugh-Nagumo(FHN)神经元在弱信号检测中的随机共振机制,实现低剂量肺部CT图像的增强。方法本文以低剂量肺部CT影像为例,在验证FHN神经元随机共振能够增强一维含噪信号有用信息的基础上,引入基于FHN神经元随机共振的图像增强方法:分别实现二维CT影像在行方向和列方向上的降维操作,以此作为FHN神经元随机共振的输入序列;FHN神经元的输出信号经过光栅扫描逆过程以及幅值映射放大处理后,通过判别器,完成低信噪比CT影像的增强。结果 FHN神经元随机共振方法增加了图像对比度,凸显影像中的有用信息。结论本文研究为图像增强技术在医学影像中的应用提供了一种有效的新思路。
Objective To study the stochastic resonance mechanism of FitzHugh-Nagumo（FHN） neurons on the weak signal detection,and realize enhancement of low doses of lung CT images.Methods Taking low doses of lung CT images as examples,and on the basis of FitzHugh-Nagumo（FHN） neurons stochastic resonance,image enhancement method was introduced,which enhanced the effective part of noisy and one-dimensional signal.First,the CT image was converted to one-dimensional signals at line direction and column direction by dimension reduction,and was taken as the input signals of FHN neuron.Then,the enhancement of the low-dose CT image was achieved after the processes of inverse raster scanning,amplitude mapping and discrimination to the output signals of FHN neuron.Results The image enhancement method of stochastic resonance based on FHN neurons stretched the contrast of images and highlighted the effective information of images.Conclusion This study provides an effective and brand-new idea for medical image enhancement.
出处
《航天医学与医学工程》
CAS
CSCD
北大核心
2012年第2期121-125,共5页
Space Medicine & Medical Engineering
基金
国家自然科学基金资助项目(60872090)
作者简介
作者简介：王海玲，女，硕士研究生，研究方向为模式识别与智能系统。E-mail：wh106062224@163．com
通讯作者：范影乐fan@hdu．edu．cn
