Method and device for extracting noise of compressed image, and method and device for reducing noise of compressed image

Patent metadata

Field	Value
Publication number	US-2023075881-A1
Application number	US-202217987582-A
Country	US
Kind code	A1
Filing date	Nov 15, 2022
Priority date	May 15, 2020
Publication date	Mar 9, 2023
Grant date	—

Abstract

Provided is a method of reducing noise of a compressed image, the method including obtaining a first image by applying a convolution filter, which sequentially performs down-convolution and up-convolution corresponding to the down-convolution, to a compressed image of an original image, obtaining a second image by subtracting the first image from the compressed image, obtaining noise comprising high-frequency information and a compressed artifact of the compressed image from the second image, obtaining a third image by removing the compressed artifact by applying a deep neural network (DNN) for removing the compressed artifact to the noise, and reconstructing the compressed image by summing the first image and the third image.

First claim

1 . A method of extracting noise of a compressed image, the method comprising:
obtaining a first image by applying a convolution filter, the convolution filter sequentially performing down-convolution and up-convolution corresponding to the down-convolution, to the compressed image of an original image; obtaining a second image by subtracting the first image from the compressed image; and obtaining from the second image, noise comprising high-frequency information of the compressed image and a compressed artifact of the compressed image.

2 . The method of claim 1 , wherein the first image comprises intermediate-frequency information of the compressed image and low-frequency information of the compressed image.

3 . The method of claim 1 , wherein the convolution filter is trained to obtain, from the compressed image, the first image comprising intermediate-frequency information of the compressed image and low-frequency information of the compressed image.

4 . The method of claim 3 , wherein the down-convolution is a convolution operation trained to reduce a size of the compressed image.

5 . The method of claim 3 , wherein the up-convolution is a convolution operation trained to increase the down-convoluted compressed image to an original size of the compressed image.

6 . The method of claim 1 , wherein the down-convolution and the up-convolution are transposed to each other.

7 . A method of reducing noise of a compressed image, the method comprising:
obtaining a first image by applying a convolution filter, the convolution sequentially performing down-convolution and up-convolution corresponding to the down-convolution, to the compressed image of an original image; obtaining a second image by subtracting the first image from the compressed image; obtaining from the second image, noise comprising high-frequency information of the compressed image and a compressed artifact of the compressed image; obtaining a third image by removing the compressed artifact by applying a deep neural network (DNN) for removing the compressed artifact, to the noise; and reconstructing the compressed image by summing the first image and the third image.

8 . The method of claim 7 , wherein the DNN for removing the compressed artifact is trained to reduce the compressed artifact with the high-frequency information of the compressed image and the compressed artifact of the compressed image as an input.

9 . The method of claim 7 , wherein the DNN for removing the compressed artifact comprises a convolution layer and an activation layer.

10 . The method of claim 9 , wherein the DNN for removing the compressed artifact further comprises a batch normalization layer.

11 . The method of claim 7 , wherein the first image comprises intermediate-frequency information of the compressed image and low-frequency information of the compressed image.

12 . The method of claim 7 , wherein the convolution filter is trained to obtain the first image comprising intermediate-frequency information of the compressed image and low-frequency information of the compressed image.

13 . The method of claim 7 , wherein the down-convolution is a convolution operation trained to reduce a size of the compressed image.

14 . The method of claim 7 , wherein the up-convolution is a convolution operation trained to increase the down-convoluted compressed image to an original size of the compressed image.

15 . An apparatus for reducing noise of a compressed image, the apparatus comprising:
a memory storing one or more instructions; and a processor configured to execute the one or more instructions stored in the memory, wherein the processor is further configured to:
obtain a first image by applying a convolution filter, the convolution filter sequentially performing down-convolution and up-convolution corresponding to the down-convolution, to a compressed image of an original image;
obtain a second image by subtracting the first image from the compressed image;
obtain from the second image, noise comprising high-frequency information of the compressed image and a compressed artifact of the compressed image;
obtain a third image by removing the compressed artifact by applying a deep neural network (DNN) for removing the compressed artifact to the noise; and
reconstruct the compressed image by summing the first image and the third image.

Assignees

• Samsung Electronics Co Ltd

Inventors

• Park Jaeyeon
• Ahn Iljun
• Lee Tammy

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06T2207/20081

  Training; Learning · CPC title

• G06T2207/20224

  Image subtraction · CPC title

• G06T2207/20084

  Artificial neural networks [ANN] · CPC title