摘要
提出了一种基于改进BP神经网络和粒子群优化算法(PSO)的图像滤波方法。该方法利用对数最小均方误差函数(LNLS)代替BP神经网络传统的最小均方误差函数(LMS),用来减小图像噪声对神经网络精度的影响;并将改进后的BP神经网络利用PSO算法优化,从而避免神经网络陷入局部极小值点,进一步提高神经网络滤波能力。实验结果表明,与传统滤波方法相比,该方法不仅能有效地滤除图像中的高斯噪声而且能很好地保护图像细节。
A new method for image noise reduction based on Particle Swarm Optimization （PSO） and modified BP neural network is proposed in this paper. This method introduces BP neural network by utilizing least mean log squares （LMLS） error function instead of least mean squares （LMS） error function as its cost function, and then the modified BP neural network optimized with PSO. The proposed method can minish the influence on the accuracy of BP neural network model which controlled by image noise and avoid local minimum obviously. Experimental results demonstrate that the proposed new method can reduce Gaussian noise of images and preserve image details more effectively than traditional algorithms.
出处
《激光杂志》
CAS
CSCD
北大核心
2009年第4期34-36,共3页
Laser Journal
基金
教育部新世纪优秀人才支持计划项目(批准号:NCET-05-0897)
关键词
图像滤波
BP神经网络
对数最小均方误差
粒子群优化算法
image filtering
BP neural network
hyperbolic tangent error function
Particle Swarm Optimization
