Dual fisheye image stitching for spherical video
AssigneeQUALCOMM Incorporated
InventorsBijan Forutanpour, Phi Hung Le Nguyen, Ning Bi
Life of this patent
Estimating the term…
Abstract
Examples are described for overlaying circular images onto a three-dimensional mesh model to generated respective portions of image content for 360-degree viewable content, where each of the portions is a sub-capsule that is larger than half of the sphere. The portions are rendered and then blended based on overlapping portions of the image content from the 360-degree viewable content.
Claims30
A method for generating image content, the method comprising:
- overlaying a first image on a first three-dimensional (3D) mesh model based on a first set of vertex coordinates for the first 3D mesh model to generate a first portion of image content;
- overlaying a second image on a second 3D mesh model based on a second set of vertex coordinates for the second 3D mesh model to generate a second portion of the image content, wherein the first portion and the second portion include overlapping image content;
- rendering the first portion and the second portion to generate a rendered first portion and a rendered second portion, respectively;
- and
- generating a sphere of image content based at least in part on blending the first rendered portion and the second rendered portion based on the overlapping image content.
A device for generating image content, the device comprising:
- a memory configured to store a first image and a second image;
- and
- a graphics processing unit (GPU) comprising at least one of fixed-function or programmable circuitry configured to:
- overlay the first image stored in the memory on a first three-dimensional (3D) mesh model based on a first set of vertex coordinates for the first 3D mesh model to generate a first portion of image content;
- overlay the second image stored in the memory on a second 3D mesh model based on a second set of vertex coordinates for the second 3D mesh model to generate a second portion of the image content, wherein the first portion and the second portion include overlapping image content;
- render the first portion and the second portion to generate a rendered first portion and a rendered second portion, respectively;
- and
- generate a sphere of image content based at least in part on blending the first rendered portion and the second rendered portion based on the overlapping image content.
A device for generating image content, the device comprising:
- means for overlaying a first image on a first three-dimensional (3D) mesh model based on a first set of vertex coordinates for the first 3D mesh model to generate a first portion of image content;
- means for overlaying a second image on a second 3D mesh model based on a second set of vertex coordinates for the second 3D mesh model to generate a second portion of the image content, wherein the first portion and the second portion include overlapping image content;
- means for rendering the first portion and the second portion to generate a rendered first portion and a rendered second portion, respectively;
- and
- means for generating a sphere of image content based at least in part on blending the first rendered portion and the second rendered portion based on the overlapping image content.
A computer-readable storage medium comprising instructions that when executed cause one or more processors of a device for generating image content to:
- overlay a first image on a first three-dimensional (3D) mesh model based on a first set of vertex coordinates for the first 3D mesh model to generate a first portion of image content;
- overlay a second image on a second 3D mesh model based on a second set of vertex coordinates for the second 3D mesh model to generate a second portion of the image content, wherein the first portion and the second portion include overlapping image content;
- render the first portion and the second portion to generate a rendered first portion and a rendered second portion, respectively;
- and
- generate a sphere of image content based at least in part on blending the first rendered portion and the second rendered portion based on the overlapping image content.
Dependencies are inferred from the claim text.