Data processing device and data processing method

1 . An oral image processing method comprising:

obtaining, as a first image, a face image of an object showing a teeth area;

obtaining, as an additional image, a face image of the same object captured in the same direction as the first image and including a greater teeth area than the first image, the greater teeth area including the teeth area shown in the first image and an additional teeth area not shown in the first image;

replacing the greater teeth area included in the additional image with a virtual teeth template to generate a modified additional image;

obtaining, as a virtual teeth area, a region corresponding to the teeth area included in the first image from the modified additional image;

obtaining a second image by replacing the teeth area included in the first image with the virtual teeth area, the virtual teeth area being a two-dimensional image region; and

using a neural network trained to learn images with different visible teeth and to generate a face image having a different teeth area from an input face image, obtaining, from the second image, a third image including a face having a different position or size of a visible teeth area compared to a face included in the second image.

2 . The oral image processing method of claim 1 , wherein the neural network is a deep neural network (DNN) for obtaining a face image from a multi-domain by converting an attribute of the input face image, wherein the DNN includes a star generative adversarial network (StarGAN).

3 . The oral image processing method of claim 1 , further comprising detecting a lip line included in the face image of the object included in the first image,

wherein the teeth area comprises an internal area of the detected lip line, and

the virtual teeth area comprises an area corresponding to the internal area of the detected lip line from the modified additional image.

4 . The oral image processing method of claim 1 , wherein the obtaining of the second image comprises:

obtaining a first feature point from the teeth area included in the first image;

obtaining a second feature point from the modified additional image; and

aligning the first feature point and the second feature point.

5 . The oral image processing method of claim 4 , wherein the obtaining of the second image further comprises receiving, from a user, selection on the first feature point and the second feature point from each of the teeth area and the modified additional image.

6 . An oral image processing apparatus comprising a processor configured to execute at least one instruction,

wherein the processor is configured to execute the at least one instruction to:

obtain, as a first image, a face image of an object showing a teeth area;

obtain, as an additional image, a face image of the same object captured in the same direction as the first image and including a greater teeth area than the first image, the greater teeth area including the teeth area shown in the first image and an additional teeth area not shown in the first image;

replace the greater teeth area included in the additional image with a virtual teeth template to generate a modified additional image;

obtain, as a virtual teeth area, a region corresponding to the teeth area included in the first image from the modified additional image;

obtain a second image by replacing the teeth area included in the first image with the virtual teeth area, the virtual teeth area being a two-dimensional image region; and

using a neural network trained to learn images with different visible teeth and to generate a face image having a different teeth area from an input face image, obtain, from the second image, a third image including a face having a different position or size of a visible teeth area compared to a face included in the second image.

7 . A non-transitory computer-readable recording medium having recorded thereon a program for implementing an oral image processing method comprising:

obtaining, as a first image, a face image of an object showing a teeth area;

obtaining, as an additional image, a face image of the same object captured in the same direction as the first image and including a greater teeth area than the first image, the greater teeth area including the teeth area shown in the first image and an additional teeth area not shown in the first image;

replacing the greater teeth area included in the additional image with a virtual teeth template to generate a modified additional image;

obtaining, as a virtual teeth area, a region corresponding to the teeth area included in the first image from the modified additional image;

obtaining a second image by replacing the teeth area included in the first image with the virtual teeth area, the virtual teeth area being a two-dimensional image region; and

using a neural network trained to learn images with different visible teeth and to generate a face image having a different teeth area from an input face image, obtaining, from the second image, a third image including a face having a different position or size of a visible teeth area compared to a face included in the second image.