摘要
应用传统方法对图像去噪处理后,图像的峰值信噪比仍旧比较低,文章提出了基于多尺度卷积神经网络的图像去噪方法。以多尺度卷积神经网络为架构,由去噪模块与边缘模块组建成多尺度卷积神经网络去噪模型,利用残差学习法对模型进行训练,并利用寻优迭代算法对代价函数进行求解,利用训练好的去噪模型对图像进行多尺度卷积计算,根据噪声真值对图像平滑处理,实现图像去噪。通过实验证明,本次设计方法去噪后图像噪声有了明显降低,峰值信噪比高于传统方法。
After image denoising with traditional methods, the peak signal to noise ratio of the image is still relatively low. This paper proposes an image denoising method based on multi-scale convolution neural network. Based on multi-scale convolutional neural network, a multi-scale convolutional neural network de-noising model is formed by de-noising module and edge module. The model is trained using residual learning method, and the cost function is solved using optimization iterative algorithm. The trained de-noising model is used to perform multi-scale convolution calculation on the image, and the image is smoothed according to the true value of the noise to achieve image de-noising. The experimental results show that the noise of the image is significantly reduced after denoising by this design method, and the peak signal to noise ratio is higher than that of traditional methods.
作者
汤勇峰
Tang Yongfeng(Jiangsu Provincial Xuzhou Pharmaceutical Vocational College,Xuzhou 221116,China)
出处
《无线互联科技》
2022年第24期154-156,共3页
Wireless Internet Technology
关键词
多尺度卷积神经网络
去噪
峰值信噪比
残差学习法
寻优迭代算法
multi-scale convolution neural network
noise elimination
peak signal to noise ratio
residual learning method
optimization iterative algorithm
