摘要
针对非局部均值去噪算法在图像去噪过程中参数选择等问题上存在的不足,给出一种基于神经网络的非局部均值去噪算法。该方法利用非局部均值去噪算法中的全局性联系,首先提取图像的非局部数据作为神经网络的输入,然后利用利用图像的非局部数据训练神经网络,最后用训练好的神经网络对噪声图像进行滤波。该算法能够减少传统非局部均值去噪算法的参数选择过程,降低了算法的复杂度。实验结果表明,与传统的非局部均值去噪算法相比,本文算法在视觉质量实验、峰值信噪比实验以及结构相似性实验上均有更好的结果。
Aiming at the shortcomings of the non-local means denoising algorithm in parameter selection during image denoising, a non-local means denoising algorithm is presented based on neural network. The method utilizes global connections in a non-local means denoising algorithm, first, the non-local data of the image is extracted as the input of the neural network, and then the neural network is trained by using the non-local data of the image, and finally the noise image is filtered by the trained neural network. The algorithm can reduce the parameter selection process of the traditional non-local means denoising algorithm and reduce the complexity of the algorithm. The experimental results show that compared with the traditional non-local means denoising algorithm, the proposed algorithm performs better in visual quality experiments, peak signal-to-noise ratio experiments and structural similarity experiments.
作者
闫涵
张旭秀
丁鸣艳
Yan Han;Zhang Xuxiu;Ding Mingyan(Dalian Jiaotong University,Dalian 116021,China)
出处
《电子测量技术》
2019年第22期145-149,共5页
Electronic Measurement Technology
基金
国家留学基金委资助计划(201608210308)
国家支撑计划(2015BAF20B02)
国家自然科学基金(61471080/F010408)
辽宁省自然基金指导计划(1553737612631)资助
关键词
图像去噪
非局部均值去噪
高斯噪声
神经网络
低照度
image denoising
non-local means algorithm
Gaussian noise
neural network
low illumination
