Apparatus and method with artificial intelligence for scaling image data
The disclosure relates to an artificial intelligence (AI) system that uses a machine learning algorithm and an application thereof. A method for controlling an electronic apparatus according to the disclosure includes receiving image data and information associated with a filter set that is applied to an artificial intelligence model for upscaling the image data from an external server; decoding the image data; upscaling the decoded image data using a first artificial intelligence model that is obtained based on the information associated with the filter set; and providing the upscaled image data for output.
1. A server comprising:
a memory; and
a processor configured to:
obtain downscaled image data by downscaling original image data;
based on a multi-artificial intelligence (AI) filter option being turned on, obtain a plurality of upscaled image data by respectively inputting the downscaled image data into a plurality of artificial intelligence (AI) upscaling models stored in the memory, identify a filter index among a plurality of filter indexes of the plurality of AI upscaling models by comparing the original image data and each of the plurality of upscaled image data, and encode the downscaled image data based on the identified filter index, the identified filter index being a filter index of an AI upscaling model among the plurality of AI upscaling models; and
based on the multi-AI filter option not being turned on, encode the downscaled image data based on a value indicating that no filter index is used.
2. The server of claim 1 , the processor is further configured to:
obtain a plurality of differences by comparing the original image data and the each of the plurality of upscaled image data, and
identify the filter index of the AI upscaling model corresponding to a minimum difference among the plurality of differences.
3. The server of claim 1 , further comprising:
a communication interface comprising communication circuitry,
wherein the processor is further configured to control the communication interface to transmit the encoded image data to an electronic apparatus.
4. The server of claim 1 , wherein each of the plurality of AI upscaling models comprises a plurality of filter sets.
5. The server of claim 4 , wherein parameters of the plurality of filter sets are trained to reduce a difference between the plurality of upscaled image data and the original image data.
6. The server of claim 1 , wherein the processor is further configured to obtain the downscaled image data by inputting the original image data into an AI downscaling model for downscaling image data.
7. The server of claim 6 , wherein a number of filters of the AI upscaling model is smaller than a number of filters of the AI downscaling model.
8. The server of claim 1 , wherein the plurality of AI upscaling models are a Convolutional Neural Network (CNN).
9. A method for controlling a server, the method comprising:
obtaining downscaled image data by downscaling original image data;
based on a multi-artificial intelligence (AI) filter option being turned on, obtaining a plurality of upscaled image data by respectively inputting the downscaled image data into a plurality of artificial intelligence (AI) upscaling models;, identifying a filter index among a plurality of filter indexes of the plurality of AI upscaling models by comparing the original image data and each of the plurality of upscaled image data, and encoding the downscaled image data based on the identified filter index, the identified filter index being a filter index of an AI upscaling model among the plurality of AI upscaling models; and
based on the multi-AI filter option not being turned on, encoding the downscaled image data based on a value indicating that no filter index is used.
10. The method of claim 9 , the identifying comprises:
obtaining a plurality of differences by comparing the original image data and the each of the plurality of upscaled image data, and
identifying the filter index of the AI upscaling model corresponding to a minimum difference among the plurality of differences.
11. The method of claim 9 , further comprising:
transmitting the encoded image data to an electronic apparatus.
12. The method of claim 9 , wherein each of the plurality of AI upscaling models comprises a plurality of filter sets.
13. The method of claim 12 , wherein parameters of the plurality of filter sets are trained to reduce a difference between the plurality of upscaled image data and the original image data.
14. The method of claim 9 , wherein the obtaining the downscaled image data comprises obtaining the downscaled image data by inputting the original image data into an AI downscaling model for downscaling image data.
15. The method of claim 14 , wherein a number of filters of the AI upscaling model is smaller than a number of filters of the AI downscaling model.
16. The method of claim 9 , wherein the plurality of AI upscaling models are a Convolutional Neural Network (CNN).