Mesh generating apparatus, method and computer-readable medium, and image processing apparatus, method and computer-readable medium
AssigneeSamsung Electronics Co., Ltd.
InventorsHwa Sup Lim, Yong Sun Kim, Seung Kyu Lee, Kee Chang Lee, Ouk Choi
Life of this patent
Estimating the term…
Abstract
Disclosed is a mesh generating apparatus, method and computer-readable medium, and an image processing apparatus, method and computer-readable medium. A single 3D mesh and a single 3D object model may be generated using a single color image and a single depth image. A plurality of viewpoint images seen from a plurality of viewpoints may be generated from a single 3D object model.
Claims27
An image processing apparatus, comprising:
- a mesh generator to generate a three-dimensional (3D) mesh using two-dimensional (2D) vertexes of a color image and 3D vertexes of a depth image;
- and
- a 3D object model generator to generate a 3D object model based on the 3D mesh and the color image.
A mesh generating apparatus, comprising:
- a first processor to generate two-dimensional (2D) vertexes from a color image, and to generate a 2D mesh using the generated 2D vertexes;
- a second processor to generate three-dimensional (3D) vertexes from a depth image;
- and
- a 3D mesh generator to match the 2D mesh with the 3D vertexes and to generate a 3D mesh.
An image processing method, comprising:
- generating, by at least one processor, a three-dimensional (3D) mesh using two-dimensional (2D) vertexes of a color image and 3D vertexes of a depth image;
- and
- generating, by the at least one processor, a 3D object model based on the 3D mesh and the color image.
A mesh generating method, comprising:
- generating, by at least one processor, two-dimensional (2D) vertexes from a color image, and generating a 2D mesh using the generated 2D vertexes;
- generating, by the at least one processor, three-dimensional (3D) vertexes from a depth image;
- and
- matching, by the at least one processor, the 2D mesh with the 3D vertexes and generating a 3D mesh.
At least one non-transitory computer readable recording medium comprising computer readable instructions that control at least one processor to implement a method, comprising:
- generating a three-dimensional (3D) mesh using two-dimensional (2D) vertexes of a color image and 3D vertexes of a depth image;
- and
- generating a 3D object model based on the 3D mesh and the color image.
A method, comprising:
- generating, by at least one processor, a multi-view image, the multi-view image generated by a method, comprising:
- generating a three-dimensional (3D) mesh using two-dimensional (2D) vertexes of a color image and 3D vertexes of a depth image and generating a 3D object model based on the 3D mesh and the color image,
- inpainting occlusion areas, the occlusion areas generated in the generated 3D object model by changes in a viewpoint of a virtual camera, and
- generating the multi-view image corresponding to multiple views from the changes in the viewpoint of the virtual camera using the generated 3D object model and the inpainted occlusion areas.
A method, comprising:
- processing, by at least one processor, an image, by extracting color information to form a two-dimensional mesh within the image;
- processing, by the at least one processor, the image, by adding depth features to the image responsive to the extracted color information;
- and
- determining, by the at least one processor, a three-dimensional mesh based on the two-dimensional mesh and the depth features to build a three-dimensional object, the three-dimensional object appearing differently from multiple viewpoints based on inpainted occlusion areas.
Dependencies are inferred from the claim text.