Image denoising method and apparatus

Patent metadata

Field	Value
Publication number	US-12062158-B2
Application number	US-202117462176-A
Country	US
Kind code	B2
Filing date	Aug 31, 2021
Priority date	Mar 1, 2019
Publication date	Aug 13, 2024
Grant date	Aug 13, 2024

Abstract

This application provides an image denoising method and apparatus, and relates to the artificial intelligence field and specifically relates to the computer vision field. The method includes: performing resolution reduction processing on a to-be-processed image to obtain a plurality of images whose resolutions are lower than that of the to-be-processed image; extracting an image feature of a higher-resolution image based on an image feature of a lower-resolution image to obtain an image feature of the to-be-processed image; and performing denoising processing on the to-be-processed image based on the image feature of the to-be-processed image to obtain a denoised image. This application can improve an image denoising effect.

First claim

What is claimed is:

1. An image denoising method, comprising:
obtaining K images based on a to-be-processed image, wherein the K images are images obtained by reducing a resolution of the to-be-processed image, K is a positive integer, the K images comprise a first image to a K th image, and the to-be-processed image is a (K+1) th image;
obtaining an image feature of the to-be-processed image based on the K images, wherein an image feature of an (i+1) th image is extracted based on an image feature of an i th image, a resolution of the (i+1) th image is higher than that of the i th image, the first image to the (K+1) th image comprise the i th image and the (i+1) th image, and i is a positive integer less than or equal to K; and
performing denoising processing on the to-be-processed image based on the image feature of the to-be-processed image to obtain a denoised image, the denoising processing comprising:
performing convolution processing on the image feature of the to-be-processed image to obtain a residual estimated value of the to-be-processed image; and
superimposing the residual estimated value of the to-be-processed image on the to-be-processed image to obtain the denoised image.

2. The method according to claim 1 , wherein that an image feature of an (i+1) th image is extracted based on an image feature of an i th image comprises:
performing convolution processing on the (i+1) th image by using a first convolutional layer to an n th convolutional layer in N convolutional layers, to obtain an initial image feature of the (i+1) th image, wherein both n and N are positive integers, n is less than or equal to N, and N is a total quantity of convolutional layers used when the image feature of the (i+1) th image is extracted;
fusing the initial image feature of the (i+1) th image with the image feature of the i th image to obtain a fused image feature; and
performing convolution processing on the fused image feature by using an (n+1) th convolutional layer to an N th convolutional layer in the N convolutional layers, to obtain the image feature of the (i+1) th image.

3. The method according to claim 1 , wherein the obtaining K images based on a to-be-processed image comprises:
performing shuffle operations on the to-be-processed image for K times, to obtain the first image to the K th image whose resolutions and channel quantities are different from those of the to-be-processed image, wherein the resolution of the i th image is lower than that of the to-be-processed image, and a channel quantity of the i th image is determined based on the channel quantity of the to-be-processed image and a ratio of the resolution of the i th image to the resolution of the to-be-processed image.

4. The method according to claim 3 , wherein a ratio of the channel quantity of the i th image to the channel quantity of the to-be-processed image is less than or equal to the ratio of the resolution of the i th image to the resolution of the to-be-processed image.

5. An image denoising apparatus, comprising a processor and a receiving interface, the processor is configured to execute one or more instructions to cause the apparatus to:
obtain K images based on a to-be-processed image, wherein the K images are images obtained by reducing a resolution of the to-be-processed image, K is a positive integer, the K images comprise a first image to a K th image, and the to-be-processed image is a (K+1) th image;
obtain an image feature of the to-be-processed image based on the K images, wherein an image feature of an (i+1) th image is extracted based on an image feature of an i th image, a resolution of the (i+1) th image is higher than that of the i th image, the first image to the (K+1) th image comprise the i th image and the (i+1) th image, and i is a positive integer less than or equal to K; and
perform denoising processing on the to-be-processed image based on the image feature of the to-be-processed image to obtain a denoised image, the denoising processing comprising:
performing convolution processing on the image feature of the to-be-processed image to obtain a residual estimated value of the to-be-processed image; and
superimposing the residual estimated value of the to-be-processed image on the to-be-processed image to obtain the denoised image.

6. The apparatus according to claim 5 , wherein the processing unit is further configured to:
perform convolution processing on the (i+1) th image by using a first convolutional layer to an n th convolutional layer in N convolutional layers, to obtain an initial image feature of the (i+1) th image, wherein both n and N are positive integers, n is less than or equal to N, and N is a total quantity of convolutional layers used when the image feature of the (i+1) th image is extracted;
fuse the initial image feature of the (i+1) th image with the image feature of the i th image to obtain a fused image feature; and
perform convolution processing on the fused image feature by using an (n+1) th convolutional layer to an N th convolutional layer in the N convolutional layers, to obtain the image feature of the (i+1) th image.

7. The apparatus according to claim 5 , wherein the processing unit is further configured to:
perform shuffle operations on the to-be-processed image for K times, to obtain the first image to the K th image whose resolutions and channel quantities are different from those of the to-be-processed image, wherein the resolution of the i th image is lower than that of the to-be-processed image, and a channel quantity of the i th image is determined based on the channel quantity of the to-be-processed image and a ratio of the resolution of the i th image to the resolution of the to-be-processed image.

8. The apparatus according to claim 7 , wherein a ratio of the channel quantity of the i th image to the channel quantity of the to-be-processed image is less than or equal to the ratio of the resolution of the i th image to the resolution of the to-be-processed image.

9. A non-transitory computer-readable storage medium, wherein the computer storage medium stores a computer program, the computer program comprises a program instruction, and when the program instruction is executed by a processor, the processor is enabled to perform the method comprising:
obtaining K images based on a to-be-processed image, wherein the K images are images obtained by reducing a resolution of the to-be-processed image, K is a positive integer, the K images comprise a first image to a K th image, and the to-be-processed image is a (K+1) th image;
obtaining an image feature of the to-be-processed image based on the K images, wherein an image feature of an (i+1) th image is extracted based on an image feature of an i th image, a resolution of the (i+1) th image is higher than that of the i th image, the first image to the (K+1) th image comprise the i th image and the (i+1) th image, and i is a positive integer less than or equal to K; and
performing denoising processing on the to-be-processed image based on the image feature of the to-be-processed image to obtain a denoised image, the denoising processing comprising:
performing convolution processing on the image feature of the to-be-processed image to obtain a residual estimated value of the to-be-processed image; and
superimposing the residual estimated value of the to-be-processed image on the to-be-processed image to obtain the denoised image.

10. The non-transitory computer-readable storage medium according to claim 9 , wherein that an image feature of an (i+1) th image is extracted based on an image feature of an i th image comprises:
performing convolution processing on

Assignees

• Huawei Tech Co Ltd

Inventors

• Song Fenglong
• Liu Liu
• Wang Tao

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/045 Primary

  Combinations of networks · CPC title

• G06T5/50

  using two or more images, e.g. averaging or subtraction · CPC title

• G06T5/60

  using machine learning, e.g. neural networks · CPC title