Artificial intelligence encoding and artificial intelligence decoding methods and apparatuses using deep neural network
Provided is an artificial intelligence (AI) encoding apparatus including a memory storing one or more instructions, and a processor configured to execute the one or more instructions stored in the memory to obtain a first image by performing AI down-scaling on an original image through a deep neural network (DNN) for down-scaling, obtain artifact information indicating an artifact region in the first image, perform post-processing to change a pixel value of a pixel in the first image, based on the artifact information, and obtain image data corresponding to a result of encoding of the post-processed first image, and AI data including the artifact information.
1. A server for providing an image by using an artificial intelligence (AI), the server comprising:
one or more processors configured to execute one or more instructions stored in the server to:
select neural network (NN) setting information from a plurality of NN setting information that is pre-stored in the server;
obtain, by a down-scaling NN, a first image by performing AI down-scaling on an original image, the down-scaling NN being set with the selected NN setting information;
obtain artifact information indicating an artifact region in the first image,
change a pixel value of a pixel in the first image, based on the artifact information;
encode the first image of which the pixel value is changed to obtain image data; and
transmit, to an electronic device, the image data and AI data related to the AI down-scaling, the AI data comprising the artifact information and being used to select NN setting information from a plurality of NN setting information that is pre-stored in the electronic device.
2. The server of claim 1 , wherein the artifact information comprises an artifact map having a certain size.
3. The server of claim 2 , wherein the processor is further configured to execute the one or more instructions to:
determine whether a pixel variance per block unit of the original image and a pixel variance per block unit of the first image satisfy a certain criterion, and
obtain the artifact map having a predetermined pixel value per block unit, based on a result of the determination.
4. The server of claim 3 , wherein, among pixels comprised in the artifact map, pixels in a block unit satisfying the certain criterion have a first pixel value and pixels in a block unit not satisfying the certain criterion have a second pixel value.
5. The server of claim 4 , wherein the processor is further configured to execute the one or more instructions to:
determine an edge region in the first image, and
change the first pixel value of a region in the artifact map corresponding to the determined edge region, to the second pixel value.
6. The server of claim 2 , wherein the processor is further configured to execute the one or more instructions to morphology-process the artifact map.
7. The server of claim 3 , wherein the certain criterion is based on a result of comparing a ratio between the pixel variance per block unit of the original image and the pixel variance per block unit of the first image, to a certain value.
8. The server of claim 1 ,
wherein the one or more processors are further configured to execute the one or more instructions to change the pixel value by applying random noise to the artifact region in the first image, based on the artifact information.
9. The server of claim 8 , wherein the one or more processors are further configured to execute the one or more instructions to change the pixel value by filtering the first image to which the random noise is applied.
10. The server of claim 1 , wherein the one or more processors are further configured to execute the one or more instructions to determine a range of a random noise value, and
to change the pixel value by applying random noise included in the determined range of the random noise value, to the artifact region in the first image.
11. The server of claim 1 , wherein the processor is further configured to execute the one or more instructions to:
input the first image to an artifact detection network, and
obtain the artifact information output from the artifact detection network.
12. A method for providing an image by a server configured to use an artificial intelligence (AI), the method comprising:
obtaining a first image by performing AI down-scaling on an original image through a neural network (NN) for down-scaling;
obtaining artifact information indicating an artifact region in the first image;
changing a pixel value of a pixel in the first image, based on the artifact information;
encode the first image of which the pixel value is changed to obtain image data; and
transmit, to an electronic device, the image data and AI data related to the AI down-scaling, the AI data comprising the artifact information and being used to select NN setting information from a plurality of NN setting information that is pre-stored in the electronic device.
13. A non-transitory computer-readable recording medium having recorded thereon a program which, when executed by a processor of the server, performs the method of claim 12 .