Image de-noising method based on compression convolutional neural network

Abstract

The invention discloses an image de-noising method based on a compression convolutional neural network, including the following steps: constructing a training data set; constructing a compression de-noising convolutional neural network model; using the training data set to train the network model; and inputting a noisy image to the trained network, and subtracting the output image of the network from the noisy image to get a clear de-noised image. The de-noising convolutional neural network in the invention is mainly characterized in that the convolution layer of the original de-noising convolutional neural network is replaced with a convolution layer decomposed and compressed by a low-rank matrix. By improving the existing de-noising convolutional neural network DnCNN, the network parameters are reduced by at least 75%, the network is simplified, and an excellent de-noising effect is maintained.

Description

technical field [0001] The invention relates to the field of digital images, in particular to an image denoising method based on a compressed convolutional neural network. Background technique [0002] In terms of image denoising technology, there are traditional denoising methods, and there are also emerging methods of denoising using deep convolutional neural networks. The present invention is based on two background technologies: 1. the latest denoising convolutional neural network DnCNN, this network utilizes the convolutional neural network of about 20 or 17 layers of depth to denoise Gaussian additive noise, according to the literature of DnCNN (Zhang K ,Zuo W,Chen Y,et al.Beyond a Gaussian Denoiser:Residual Learning of Deep CNN for Image Denoising[J].arXiv preprint arXiv:1608.03981,2016) pointed out that this method can achieve the best denoising level at present , but the network has a huge amount of parameters and requires high hardware. 2. A network compression t...

AI Technical Summary

Problems solved by technology

Although DnCNN has achieved better denoising results, it also pays more space cost compared with traditional denoising methods, such as filtering based on spatial domain and filtering denoising based on transform domain.

Images