Image denoising method based on attention and dense connection residual block convolution kernel neural network

Abstract

The invention discloses an image denoising method based on an attention and dense connection residual block convolutional neural network. The method comprises the following steps: constructing a training data set, and performing an preprocessing operation on the training data set; constructing a network denoising model by using a convolutional neural network combining an attention mechanism and a dense connection residual block; setting a hyper-parameter and a loss function of the network denoising model, and optimizing the loss function; selecting images of different noise levels in the training data set, and training the network denoising model to obtain a trained network model; and performing image denoising according to the trained network model, and evaluating the noise images by using peak signal-to-noise ratio indexes. The method has the beneficial effects of improving the denoising performance and the imaging quality.

Description

technical field [0001] The invention belongs to the field of computer vision and image processing, in particular to an image denoising method of attention and densely connected residual block convolutional neural network. Background technique [0002] Images are often disturbed by many adverse factors during the process of acquisition, storage, recording and transmission, which degrades, distorts and degrades the quality to a certain extent, resulting in the existence of noise in the acquired images, thus affecting the quality of the images. Therefore, in order to obtain high-quality digital images and restore the information of the original image from the noisy image, it is necessary to perform noise reduction processing on the image to maintain the integrity of the original information as much as possible while removing useless information in the signal. for subsequent applications. [0003] Image denoising is a classic problem in the field of image processing, and it is ...

AI Technical Summary

Problems solved by technology

Traditional image denoising methods, the more common ones include three-dimensional filtering based on non-local block matching (BM3D), weighted kernel norm minimization (WNNM) denoising algorithms, although these algorithms can remove noise in images, but these methods During the testing process, parameters need to be manually selected, which involves complex optimization problems and takes a lot of time and cost

Images