Electronic device and operation method thereof
A method of an electronic device including obtaining a low-resolution input image by down-sampling a high-resolution input image; obtaining a low-resolution output image by performing image quality processing on the low-resolution input image; obtaining a low-resolution model from a conversion relationship between the low-resolution input image prior to the image quality processing being performed and the low-resolution output image subsequent to the image quality processing being performed; performing up-sampling of the low-resolution model; obtaining a high-resolution model by modifying the up-sampled low-resolution model, based on a difference between the high-resolution input image and the low-resolution input image; and obtaining a high-resolution output image from the high-resolution input image, by applying the high-resolution model to the high-resolution input image.
1. A method of an electronic device, the method comprising:
obtaining a low-resolution input image by down-sampling a high-resolution input image;
obtaining a low-resolution output image by performing image quality processing on the low-resolution input image;
obtaining a low-resolution model from a conversion relationship between the low-resolution input image prior to the image quality processing being performed and the low-resolution output image subsequent to the image quality processing being performed;
performing up-sampling of the low-resolution model;
obtaining a high-resolution model by modifying the up-sampled low-resolution model, based on a difference between the high-resolution input image and the low-resolution input image; and
obtaining a high-resolution output image from the high-resolution input image, by applying the high-resolution model to the high-resolution input image,
wherein the down-sampling of the high-resolution input image comprises dividing the high-resolution input image into a plurality of pixel groups and selecting a pixel to be sampled independently for each of the plurality of pixel groups.
2. The method of claim 1 , wherein the down-sampling of the high-resolution input image comprises selecting a pixel having a pixel value change amount in the plurality of pixel groups equal to or greater than a reference value, for each of the plurality of pixel groups.
3. The method of claim 1 , wherein the obtaining of the low-resolution model comprises obtaining, as the low-resolution model, a linear regression model indicating a conversion relationship between a pixel value of an input pixel of the low-resolution input image and a pixel value of an output pixel of the low-resolution output image, the output pixel corresponding to the input pixel.
4. The method of claim 3 , wherein the linear regression model is obtained by using a conversion relationship between pixel values of a reference region including the input pixel and pixel values of a region corresponding to the reference region and including the output pixel.
5. The method of claim 4 , wherein the obtaining of the low-resolution model further comprises:
obtaining pixel value change amounts of pixels included in a standard region in the low-resolution input image, the standard region including the input pixel; and
obtaining the reference region with an adjusted size by adjusting a size of the standard region according to the pixel value change amounts.
6. The method of claim 3 , wherein the up-sampling of the low-resolution model comprises obtaining a linear regression model regarding pixels of the high-resolution input image down-sampled to the input pixel by using the linear regression model regarding the input pixel.
7. The method of claim 3 , wherein the input pixel is a first input pixel and the up-sampling of the low-resolution model comprises:
obtaining a linear regression model regarding pixels of the high-resolution input image down-sampled to the first input pixel, by interpolating the linear regression model regarding the first input pixel and a linear regression model regarding a second input pixel adjacent to the first input pixel.
8. The method of claim 1 , wherein the obtaining of the high-resolution model comprises:
obtaining a reliability weight for each pixel of the high-resolution input image, based on a pixel value difference between the high-resolution input image and the low-resolution input image; and
obtaining the high-resolution model by applying the reliability weight to the up-sampled low-resolution model.
9. The method of claim 8 , wherein the reliability weight is obtained based on a pixel value difference between a certain region of the high-resolution input image and the low-resolution input image obtained by being down-sampled from the certain region.
10. An electronic device comprising:
a signal processor;
a memory storing one or more instructions; and
a processor configured to execute the one or more instructions stored in the memory to control the signal processor,
wherein the signal processor is configured to, according to control by the processor:
obtain a low-resolution input image by down-sampling a high-resolution input image;
obtain a low-resolution output image by performing image quality processing on the low-resolution input image;
obtain a low-resolution model from a conversion relationship between the low-resolution input image prior to the image quality processing being performed and the low-resolution output image subsequent to the image quality processing being performed;
perform up-sampling of the low-resolution model;
obtain a high-resolution model by modifying the up-sampled low-resolution model, based on a difference between the high-resolution input image and the low-resolution input image; and
obtain a high-resolution output image from the high-resolution input image, by applying the high-resolution model to the high-resolution input image,
wherein the signal processor is further configured to divide the high-resolution input image into a plurality of pixel groups and down-sample the high-resolution input image by selecting a pixel to be sampled independently for each of the plurality of pixel groups.
11. The electronic device of claim 10 , wherein the signal processor is further configured to down-sample the high-resolution input image by selecting a pixel having a pixel value change amount in the plurality of pixel groups equal to or greater than a reference value, for each of the plurality of pixel groups.
12. The electronic device of claim 10 , wherein the signal processor is further configured to obtain as the low-resolution model, a linear regression model indicating a conversion relationship between a pixel value of an input pixel of the low-resolution input image and a pixel value of an output pixel of the low-resolution output image, the output pixel corresponding to the input pixel.
13. The electronic device of claim 12 , wherein the signal processor is further configured to obtain the linear regression model by using a conversion relationship between pixel values of a reference region including the input pixel and pixel values of a region corresponding to the reference region and including the output pixel.
14. The electronic device of claim 13 , wherein the signal processor is further configured to:
obtain pixel value change amounts of pixels included in a standard region in the low-resolution input image, the standard region including the input pixel; and
obtain the reference region with an adjusted size by adjusting a size of the standard region according to the pixel value change amounts.
15. The electronic device of claim 12 , wherein the signal processor is further configured to up-sampling the low-resolution model by obtaining a linear regression model regarding pixels of the high-resolution input image down-sampled to the input pixel by using the linear regression model regarding the input pixel.
16. The electronic device of claim 12 , wherein the input pixel is a first input pixel and the signal processor is further configured to:
obtain a linear regression model regarding pixels of the high-resolution input image down-sampled to the first input pixel, by interpolating the linear regression model regarding the first input pixel and a linear regression model regarding a second input pixel adjacent to the first input pixel.
17. The electronic device of claim 10 , wherein the signal processor is further configured to obtain a reliability weight for each pixel of a certain region of the high-resolution input image, based on a pixel value difference between the certain region and the low-resolution input image obtained by being down-sampled from the certain region, and obtain the high-resolution model by applying the reliability weight to the up-sampled low-resolution model.
18. A non-transitory computer-readable recording medium having recorded thereon a program for implementing an operation of an electronic device, the operation comprising:
obtaining a low-resolution input image by down-sampling a high-resolution input image;
obtaining a low-resolution output image by performing image quality processing on the low-resolution input image;
obtaining a low-resolution model from a conversion relationship between the low-resolution input image prior to the image quality processing being performed and the low-resolution output image subsequent to the image quality processing being performed;
performing up-sampling of the low-resolution model;
obtaining a high-resolution model by modifying the up-sampled low-resolution model, based on a difference between the high-resolution input image and the low-resolution input image; and
obtaining a high-resolution output image from the high-resolution input image by applying the high-resolution model to the high-resolution input image,
wherein the down-sampling of the high-resolution input image comprises dividing the high-resolution input image into a plurality of pixel groups and selecting a pixel to be sampled independently for each of the plurality of pixel groups.