A neural network-based image denoising method and system

Abstract

The invention discloses an image denoising method based on a neural network, comprising: acquiring a high-sensitivity image and a low-sensitivity image of a scene; generating a first neural network model, and using the first neural network model to conduct Denoising processing to obtain denoised low-sensitivity images; high-sensitivity images and denoised low-sensitivity images constitute a training image database, and use the training image database to train neural network models to generate images for image denoising The second neural network model: the image to be processed is used as an input of the second neural network model, and a corresponding clean image is obtained after processing. In addition, a neural network-based image denoising system is also disclosed. The above neural network-based image denoising method and system can process images with real noise information, and has practical application significance.

Description

technical field [0001] The invention relates to the fields of mathematical image processing and pattern recognition, in particular to an image denoising method and system based on a neural network. Background technique [0002] Use the method of pattern recognition to filter the noise image to obtain the noise corresponding to the noise image, and then use the image pixel correspondence to separate the corresponding noise from the noise image to obtain the corresponding clean image. [0003] In traditional image denoising processing, pattern recognition methods including CNN (Convolution Neural Network Convolutional Neural Network) have important applications, but the training image databases used to construct these convolutional neural networks are mostly images containing artificial noise database, so it can only deal with artificially synthesized noise, and has little effect on real noisy images. [0004] Considering that the noise in real life is polygonal, rather than ...

AI Technical Summary

Problems solved by technology

Although the convolutional neural network constructed using the above training image database can process the actual noise during image denoising processing, in practical applications, the clean image obtained after processing still does not have high-quality characteristics.

It can be seen that the image denoising method in the prior art cannot achieve a good denoising effect

Images