US 12,034,906 B2
Methods, systems, and computer-readable storage media for generating three-dimensional (3D) images of a scene
Michael McNamer, Apex, NC (US); Marshall Robers, Durham, NC (US); Tassos Markas, Chapel Hill, NC (US); and Jason Hurst, Cary, NC (US)
Assigned to 3DMedia Corporation, Chapel Hill, NC (US)
Filed by Michael McNamer, Apex, NC (US); Marshall Robers, Durham, NC (US); Tassos Markas, Chapel Hill, NC (US); and Jason Hurst, Cary, NC (US)
Filed on Jun. 18, 2021, as Appl. No. 17/351,609.
Application 17/351,609 is a continuation of application No. 16/129,273, filed on Sep. 12, 2018, granted, now 11,044,458.
Application 16/129,273 is a continuation of application No. 15/193,623, filed on Jun. 27, 2016, granted, now 10,080,012, issued on Sep. 18, 2018.
Application 15/193,623 is a continuation of application No. 13/115,459, filed on May 25, 2011, granted, now 9,380,292, issued on Jun. 28, 2016.
Application 13/115,459 is a continuation in part of application No. 12/842,084, filed on Jul. 23, 2010, granted, now 8,508,580, issued on Aug. 13, 2013.
Application 13/115,459 is a continuation in part of application No. 12/842,171, filed on Jul. 23, 2010, granted, now 8,436,893, issued on May 7, 2013.
Claims priority of provisional application 61/230,133, filed on Jul. 31, 2009.
Claims priority of provisional application 61/230,131, filed on Jul. 31, 2009.
Prior Publication US 2021/0314547 A1, Oct. 7, 2021
Int. Cl. H04N 13/221 (2018.01); G06T 7/593 (2017.01); H04N 13/257 (2018.01); H04N 13/261 (2018.01); H04N 23/63 (2023.01)
CPC H04N 13/221 (2018.05) [G06T 7/593 (2017.01); G06T 7/596 (2017.01); H04N 13/257 (2018.05); H04N 13/261 (2018.05); H04N 23/63 (2023.01)] 16 Claims
OG exemplary drawing
 
1. An apparatus for continuously determining three-dimensional images from a plurality of images, the apparatus comprising:
receiving a plurality of images;
select two images from the plurality of the received images;
perform a series of computations between the selected two images to generate a plurality of image characteristics;
compare the generated characteristics between the two images to determine whether the two images are similar given a set of criteria for each characteristic;
if one or more of the similarity criteria is met, the two images are considered as a stereoscopic candidate pair;
the stereoscopic candidate pair is further analyzed to determine whether the two images can be used to create a stereoscopic pair by performing the following operations:
perform edge detection on the stereoscopic candidate pair;
match detected edges between the two images and determine image correspondence;
in case there is image correspondence, determine image that corresponds to the left eye and which image corresponds to the right eye;
rectify the two images to correct for camera viewpoint;
compute and adjust parallax of the rectified images;
create a stereoscopic pair based on the parallax adjusted image.