Apparatus and method for generating image

1. An image generating apparatus comprising:

a display configured to output an image;

a memory configured to store one or more instructions; and

a processor configured to execute the one or more instructions to:

detect an object in an input image including a plurality of frames;

provide a plurality of candidate boundaries for masking the detected object;

identify a boundary of the detected object by assessing the plurality of candidate boundaries;

mask the object in one of the plurality of frames by using the identified boundary; and

generate a partial moving image with the object moving by using the one of the plurality of frames in which the object is masked and the plurality of frames,

wherein each of the plurality of candidate boundaries is a candidate boundary contour of the detected object, and the one of the plurality of frames in which the object is masked and the plurality of frames are included in the input image.

2. The image generating apparatus of claim 1 , wherein the processor is further configured to execute the one or more instructions to provide the plurality of candidate boundaries for masking the object detected in the input image by using a first artificial intelligence (AI) model.

3. The image generating apparatus of claim 2 , wherein the first AI model includes a plurality of segmentation AI models, and

the processor is further configured to execute the one or more instructions to provide the plurality of candidate boundaries by using the plurality of segmentation AI models.

4. The image generating apparatus of claim 1 , wherein the processor is further configured to execute the one or more instructions to assess the plurality of candidate boundaries and obtain a result of the assessing by using a second artificial intelligence (AI) model.

5. The image generating apparatus of claim 4 , wherein the second AI model includes at least one of an AI model for performing a technical assessment on input data or an AI model for performing an aesthetic assessment on the input data.

6. The image generating apparatus of claim 4 , further comprising a user interface,

wherein the display is further configured to output a preset number of boundaries having high ranks from among the plurality of candidate boundaries according to the result of the assessing, that is obtained by using the second AI model, and

the processor is further configured to execute the one or more instructions to identify, as the boundary, a boundary selected by a user's selection via the user interface from among the preset number of boundaries output by the display.

7. The image generating apparatus of claim 6 , wherein the second AI model receives and learns information about the user's selection regarding the boundary and uses a result of the learning when assessing the plurality of candidate boundaries.

8. The image generating apparatus of claim 7 , wherein the second AI model classifies and learns the information about the user's selection for each of a plurality of users.

9. The image generating apparatus of claim 1 , further comprising a user interface,

wherein the processor is further configured to execute the one or more instructions to:

identify objects that are in motion in the input image by tracking the plurality of frames; and

detect the object selected by a user via the user interface from among the identified objects.

10. An image generating method comprising:

detecting an object in an image including a plurality of frames;

providing a plurality of candidate boundaries for masking the detected object;

identifying a boundary by assessing the plurality of candidate boundaries;

masking the object in one of the plurality of frames by using the identified boundary; and

generating a partial moving image with the object moving by using the one of the plurality of frames in which the object is masked and the plurality of frames,

wherein each of the plurality of candidate boundaries is a candidate boundary contour of the detected object, and the one of the plurality of frames in which the object is masked and the plurality of frames are included in the input image.

11. The image generating method of claim 10 , wherein the providing the plurality of candidate boundaries further comprises:

providing the plurality of candidate boundaries for masking the object detected in the image by using a first artificial intelligence (AI) model.

12. The image generating method of claim 11 , wherein the first AI model includes a plurality of segmentation AI models, and

wherein the providing the plurality of candidate boundaries further comprises providing the plurality of candidate boundaries by using the plurality of segmentation AI models.

13. The image generating method of claim 10 , wherein the identifying the boundary further comprises:

assessing the plurality of candidate boundaries and obtaining a result of the assessing by using a second artificial intelligence (AI) model.

14. The image generating method of claim 13 , wherein the second AI model includes at least one of an AI model for performing a technical assessment on input data or an AI model for performing an aesthetic assessment on the input data, and

the assessing the plurality of candidate boundaries further comprises performing at least one of the technical assessment or the aesthetic assessment on the plurality of candidate boundaries by using the second AI model.

15. The image generating method of claim 13 , wherein the identifying the boundary further comprises:

outputting a preset number of boundaries having high ranks from among the plurality of candidate boundaries according to the result of the assessing, that is obtained by using the second AI model; and

identifying, as the boundary, a boundary selected by a user's selection from among the output preset number of boundaries.

16. The image generating method of claim 15 , wherein the assessing the plurality of candidate boundaries further comprises:

receiving and learning information about the user's selection regarding the boundary, by the second AI model; and

using a result of the learning when assessing the plurality of candidate boundaries.

17. The image generating method of claim 16 , wherein the receiving and the learning the information about the user's selection further comprises classifying and learning the information about the user's selection for each of a plurality of users.

18. A non-transitory computer-readable recording medium having recorded thereon a program which, when executed by a computer system, causes the computer system to execute an image generating method including:

detecting an object in an image including a plurality of frames;

providing a plurality of candidate boundaries for masking the detected object;

identifying a boundary by assessing the plurality of candidate boundaries;

masking the object in one of the plurality of frames by using the identified boundary; and

generating a partial moving image with the object moving by using the one of the plurality of frames in which the object is masked and the plurality of frames,

wherein each of the plurality of candidate boundaries is a candidate boundary contour of the detected object, and the one of the plurality of frames in which the object is masked and the plurality of frames are included in the input image.

19. The image generating apparatus according to claim 1 , wherein the processor is further configured to execute the one or more instructions to:

assess the plurality of candidate boundaries and obtain a result of the assessing by using an artificial intelligence (AI) model, wherein, as the result of the assessing, the AI model outputs a preset number of boundaries having high ranks among the plurality of candidate boundaries, and

identify the boundary of the detected object among the preset number of boundaries.