Method for providing XR content and XR device for providing XR content

An XR device for providing extended reality (XR) content, the XR device comprising:

• an image/video analyzer configured to:
• obtain a first two-dimensional (2D) image representing a user's face,
• determine whether an object which is not the user's face is included in the first 2D image, wherein in response to determining that the object is included in the first 2D image, the image/video analyzer is further configured to:
• generate a second 2D image representing the user's face without the object based on the first 2D image, generate a mask image representing a difference between the first 2D image and the second 2D image and determine a type of object based on the generated mask image;
• a three-dimensional (3D) image/video processor configured to:
• generate a first 3D image corresponding to the generated second 2D image, obtain a 3D preset including one or more 3D images corresponding to the determined type of object, and generate a second 3D image by combining at least one of the one or more 3D images with the first 3D image;
• a first network configured to:
• transform the second 2D image into the first 2D image again and then generate a restored image of the first 2D image through a first network;
• and
• process to minimize a cycle-consistency loss between the first 2D image and the restored image;
• and
• a display configured to providing XR content including the generated second 3D image.

An extended reality (XR) device, comprising:

• one or more processors;
• a display;
• and
• a memory storing one or more programs, wherein the one or more programs are configured to be executed by the one or more processors, the one or more programs including instructions for:
• determining whether an object which is not a user's face is included in a first two-dimensional (2D) image representing the user's face;
• in response to determining that the object is included in the first 2D image, generating a second 2D image representing the user's face without the object based on the first 2D image;
• generating a first three-dimensional (3D) image corresponding to the generated second 2D image;
• generating a mask image representing a difference between the first 2D image and the second 2D image;
• determining a type of object based on the generated mask image;
• obtaining a 3D preset including one or more 3D images corresponding to the determined type of object;
• generating a second 3D image by combining at least one of the one or more 3D images with the first 3D image, wherein the display is further configured to provide XR content including the generated second 3D image;
• transforming the second 2D image into the first 2D image again and then generating a restored image of the first 2D image through a first network;
• and
• processing to minimize a cycle-consistency loss between the first 2D image and the restored image.