Image denoising method and device, storage medium and electronic equipment

Abstract

The invention relates to an image denoising method and device, a storage medium and electronic equipment, and the method comprises the steps: receiving a to-be-processed image which comprises noise inzero-mean distribution; and inputting the to-be-processed image into an image denoising model to obtain a denoised image corresponding to the to-be-processed image, and training images adopted for training the image denoising model are all images with noise. Therefore, for the to-be-processed image including noise in zero-mean distribution, the to-be-processed image can be denoised through the image denoising model trained on the basis of the image with noise, and a target image with a high signal-to-noise ratio can be obtained without repeated scanning processing, so that the acquisition efficiency of the training image can be improved; meanwhile, the risk that other noise is possibly introduced in the target image acquisition process can be reduced, the matching degree of the input image used for training the image denoising model and the target image is ensured, the accuracy and the use range of the image denoising model are ensured, and the use experience of a user is improved.

Description

technical field [0001] The present disclosure relates to the field of image processing, and in particular, to an image denoising method, device, storage medium and electronic equipment. Background technique [0002] At present, methods for removing noise in images are mainly divided into two categories. One is denoising based on the acquisition end. This type of method mainly collects images multiple times and then averages the images to suppress noise to a certain extent and improve the image signal-to-noise ratio. The other is denoising based on the post-processing end, which includes some existing filter denoising algorithms, denoising algorithms based on statistical characteristics, and denoising algorithms based on deep learning networks. In the prior art, a denoising algorithm based on a deep learning network is usually implemented by taking a low signal-to-noise ratio image scanned once as an input, and a high signal-to-noise ratio image processed by multiple scans a...

AI Technical Summary

Problems solved by technology

However, in the above process, the process of obtaining a high signal-to-noise ratio image requires multiple scans of the image, the data acquisition efficiency is low, and it is difficult to apply to the scene where multiple scans cannot be performed.

Images