Volumetric video creation from user-generated content

Abstract

A processing system having at least one processor may obtain at least a first source video from a first endpoint device and a second source video from a second endpoint device, where each of the first source video and the second source video is a two-dimensional video, determine that the first source video and the second source video share at least one feature that is the same for both the first source video and the second source video, and generate a volumetric video from the first source video and the second source video, where the volumetric video comprises a photogrammetric combination of the first source video and the second source video.

What is claimed is: 
     
         1 . A method comprising:
 obtaining, by a processing system including at least one processor, at least a first source video from a first endpoint device and a second source video from a second endpoint device, wherein each of the first source video and the second source video is a two-dimensional video;   determining, by the processing system, that the first source video and the second source video share at least one feature that is same for both the first source video and the second source video;   identifying, by the processing system, a first object that is present in both the first source video and the second source video; and   generating, by the processing system, a volumetric video from the first source video and the second source video, wherein the volumetric video comprises a photogrammetric combination of the first source video and the second source video, and wherein the first object is excluded in the volumetric video.   
     
     
         2 . The method of  claim 1 , further comprising:
 identifying, by the processing system, a second object that is present in both the first source video and the second source video;   matching, by the processing system, the second object to a two-dimensional object that is stored in a database of two-dimensional objects; and   obtaining, by the processing system, a stored three-dimensional model that is identified, in the database of two-dimensional objects, as being matched to the two-dimensional object.   
     
     
         3 . The method of  claim 2 , wherein the matching comprises matching at least one visual feature of a two-dimensional image of the second object from at least one of the first source video or the second source video to at least one visual feature of the two-dimensional object, wherein the at least one visual feature of the two-dimensional image of the second object and the at least one visual feature of the two-dimensional object comprise at least one of:
 location information;   time information;   an event tag; or   a keyword.   
     
     
         4 . The method of  claim 3 , wherein the at least one feature comprises:
 a presence of the second object.   
     
     
         5 . The method of  claim 2 , wherein the generating further comprises:
 replacing the second object in the volumetric video with the stored three-dimensional model.   
     
     
         6 . The method of  claim 5 , wherein the generating further comprises:
 texture mapping visual information of the second object in both the first source video and the second source video to the three-dimensional model to generate an enhanced three-dimensional model of the second object; and   storing the enhanced three-dimensional model of the second object for reuse in another volumetric video.   
     
     
         7 . The method of  claim 1 , wherein the generating further comprises:
 performing, by the processing system, a bundle adjustment based upon a plurality of visual features shared by the first source video and the second source video.   
     
     
         8 . The method of  claim 7 , wherein the bundle adjustment is in accordance with a ranking of the plurality of visual features shared by the first source video and the second source video. 
     
     
         9 . The method of  claim 1 , wherein the generating the volumetric video further comprises:
 performing, by the processing system, a time alignment of the first source video and the second source video.   
     
     
         10 . The method of  claim 1 , wherein the generating the volumetric video further comprises:
 identifying, by the processing system, from the first source video and the second source video, a region of focus in a space of the volumetric video, wherein the photogrammetric combination of the first source video and the second source video is based on a prioritization of the region of focus.   
     
     
         11 . The method of  claim 10 , wherein the prioritization of the region of focus utilizes more key features from the region of focus for the photogrammetric combination than from other regions of the space of the volumetric video. 
     
     
         12 . The method of  claim 1 , further comprising:
 presenting, by the processing system, the volumetric video via at least one of:
 the first endpoint device; 
 the second endpoint device; or 
 a third endpoint device. 
   
     
     
         13 . The method of  claim 1 , further comprising:
 generating, by the processing system, an output video comprising a two-dimensional traversal of the volumetric video, wherein a perspective of the output video is different from a perspective of the first source video and a perspective of the second source video.   
     
     
         14 . The method of  claim 13 , wherein the output video is generated in accordance with a selection by a user of at least one viewing perspective within a space of the volumetric video. 
     
     
         15 . The method of  claim 14 , further comprising:
 presenting, by the processing system, the output video via at least one of:
 the first endpoint device; 
 the second endpoint device; or 
 a third endpoint device. 
   
     
     
         16 . The method of  claim 1 , further comprising:
 ranking, by the processing system, visual qualities of the first source video and the second source video; and   providing, by the processing system, at least one recommendation of at least one viewing perspective within a space of the volumetric video in accordance with the ranking of the visual qualities.   
     
     
         17 . The method of  claim 1 , further comprising:
 identifying, by the processing system, a region of a space of the volumetric video having a visual quality below a threshold;   providing, by the processing system, a recommendation to at least one of the first endpoint device or the second endpoint device to capture at least one additional source video of a region of a physical space associated with the region of the space of the volumetric video that is identified as having the visual quality below the threshold;   obtaining, by the processing system, the at least one additional source video from at least one of the first endpoint device or the second endpoint device; and   modifying, by the processing system, the volumetric video in accordance with the at least one additional source video, wherein the modifying comprises a photogrammetric combination of the first source video, the second source video, and the at least one additional source video.   
     
     
         18 . The method of  claim 17 , wherein the recommendation to capture the at least one additional source video includes a recommendation of at least one of:
 a camera position in the physical space corresponding to a position in the space of the volumetric video;   a camera orientation in the physical space corresponding to an orientation within the space of the volumetric video;   a camera zoom level; or   an object on which to focus.   
     
     
         19 . A non-transitory computer-readable medium storing instructions which, when executed by a processing system including at least one processor, cause the processing system to perform operations, the operations comprising:
 obtaining at least a first source video from a first endpoint device and a second source video from a second endpoint device, wherein each of the first source video and the second source video is a two-dimensional video;   determining that the first source video and the second source video share at least one feature that is same for both the first source video and the second source video;   identifying a first object that is present in both the first source video and the second source video; and   generating a volumetric video from the first source video and the second source video, wherein the volumetric video comprises a photogrammetric combination of the first source video and the second source video, and wherein the first object is excluded in the volumetric video.   
     
     
         20 . A device comprising:
 a processing system including at least one processor; and   a computer-readable medium storing instructions which, when executed by the processing system, cause the processing system to perform operations, the operations comprising:
 obtaining at least a first source video from a first endpoint device and a second source video from a second endpoint device, wherein each of the first source video and the second source video is a two-dimensional video; 
 determining that the first source video and the second source video share at least one feature that is same for both the first source video and the second source video; 
 identifying a first object that is present in both the first source video and the second source video; and 
 generating a volumetric video from the first source video and the second source video, wherein the volumetric video comprises a photogrammetric combination of the first source video and the second source video, and wherein the first object is excluded in the volumetric video.