Method and device for restoring image obtained from array camera

1. An image restoration method of restoring an image obtained through an array camera, the image restoration method comprising:

obtaining a plurality of images through lens elements included in the array camera;

obtaining a global parameter of the plurality of images based on a depth value of a scene obtained by the array camera;

generating first processed images by transforming a viewpoint of each of the plurality of images based on the obtained global parameter;

obtaining a local parameter for each pixel corresponding to each of the first processed images;

generating second processed images by transforming the first processed images based on the obtained local parameter; and

generating a synthesized image of a target viewpoint based on synthesizing the second processed images.

2. The image restoration method of claim 1 , wherein the obtaining of the global parameter comprises:

obtaining the global parameter based on a neural network model that receives the plurality of images obtained as an input.

3. The image restoration method of claim 2 , wherein the obtaining of the global parameter comprises:

obtaining matrix elements included in a projection matrix.

4. The image restoration method of claim 1 , wherein the generating of the synthesized image of the target viewpoint comprises:

generating the synthesized image with a higher resolution than a resolution of the second processed images by combining pixels of the second processed images.

5. The image restoration method of claim 4 , wherein the generating of the synthesized image of the target viewpoint comprises:

generating the synthesized image from the second processed images based on pixel shuffling.

6. The image restoration method of claim 1 , further comprising:

transforming the plurality of images into a plurality of higher-resolution images that have higher resolutions than resolutions of the plurality of images, respectively,

wherein the obtaining of the global parameter comprises obtaining the global parameter based on a neural network model that receives the plurality of higher-resolution images as an input.

7. The image restoration method of claim 1 , wherein the obtaining of the local parameter comprises:

obtaining an offset value of a pixel position for each pixel corresponding to each of the first processed images based on a neural network model that receives the first processed images as an input.

8. The image restoration method of claim 7 , wherein the generating of the second processed images comprises:

generating the second processed images by performing an image transformation on each pixel corresponding to each of the first processed images based on the offset value.

9. The image restoration method of claim 1 , wherein the generating of the first processed images comprises:

warping the plurality of images obtained into the first processed images having the target viewpoint based on the global parameter.

10. The image restoration method of claim 1 , wherein the lens elements of the array camera are disposed on the same plane and spaced apart from each other at a same distance.

11. The image restoration method of claim 1 , wherein the plurality of images obtained through the lens elements are viewpoint images respectively corresponding to different viewpoints.

12. A non-transitory computer-readable storage medium storing instructions that are executable by a processor to perform the image restoration method of claim 1 .

13. An image restoration device comprising:

a processor; and

a memory configured to store therein instructions to be executed by the processor,

wherein the processor is configured to:

receive a plurality of images corresponding to a plurality of viewpoints via an array camera;

estimate a global parameter of the plurality of images based on a depth value of a scene obtained by the array camera;

generate first processed images by transforming a viewpoint of the plurality of images based on the obtained global parameter;

obtain a local parameter for each pixel corresponding to each of the first processed images;

generate second processed images by transforming the first processed images based on the obtained local parameter; and

generate a synthesized image of a target viewpoint by synthesizing the second processed images.

14. The image restoration device of claim 13 , wherein the processor is further configured to:

obtain the global parameter based on a neural network model that receives the plurality of images obtained as an input.

15. The image restoration device of claim 13 , wherein the processor is further configured to:

obtain an offset value of a pixel position for each pixel corresponding to each of the first processed images based on a neural network model that receives the first processed images as an input.

16. The image restoration device of claim 13 , wherein the processor is further configured to:

generate the synthesized image with a higher resolution than a resolution of the second processed images by combining pixels of the second processed images.

17. The image restoration device of claim 13 , wherein the processor is further configured to:

transform the received images into respective higher-resolution images that have higher resolutions than resolutions of the received images, and obtain the global parameter based on a neural network model that receives the higher-resolution images as an input.

18. A mobile device comprising:

an imaging device comprising an array camera, and configured to obtain a plurality of images corresponding to a plurality of viewpoints; and

a processor configured to:

obtain a global parameter of the plurality of images based on a depth value of a scene obtained by the array camera;

generate first processed images by transforming a viewpoint of the plurality of images based on the obtained global parameter;

obtain a local parameter for each pixel corresponding to each of the first processed images;

generate second processed images by transforming the first processed images based on the obtained local parameter; and

generate a synthesized image of a target viewpoint by synthesizing the second processed images.

19. An image restoration device comprising:

a processor; and

a memory configured to store therein instructions to be executed by the processor,

wherein the processor is configured to:

receive a plurality of images corresponding to a plurality of viewpoints through lens elements included in an array camera;

obtain a global parameter of the plurality of images based on a depth value of a scene obtained by the array camera;

generate first processed images by transforming a viewpoint of the plurality of images based on the obtained global parameter;

obtain a local parameter for pixels corresponding to the first processed images, respectively, by obtaining an offset value of a position of each of the pixels based on a neural network model that receives the first processed images as an input;

generate second processed images by transforming the first processed images based on the obtained local parameter; and

generate a synthesized image of a target viewpoint by synthesizing the second processed images,

wherein the plurality of images obtained through the lens elements are viewpoint images respectively corresponding to different viewpoints.