US 12,147,505 B1
Hand pose estimation for machine learning based gesture recognition
Jonathan Marsden, San Mateo, CA (US); Raffi Bedikian, San Francisco, CA (US); and David Samuel Holz, San Francisco, CA (US)
Assigned to ULTRAHAPTICS IP TWO LIMITED, Bristol (GB)
Filed by ULTRAHAPTICS IP TWO LIMITED, Bristol (GB)
Filed on Jul. 20, 2023, as Appl. No. 18/224,373.
Application 18/224,373 is a continuation of application No. 16/508,231, filed on Jul. 10, 2019, granted, now 11,714,880.
Application 16/508,231 is a continuation of application No. 15/432,872, filed on Feb. 14, 2017, abandoned.
Claims priority of provisional application 62/335,534, filed on May 12, 2016.
Claims priority of provisional application 62/296,561, filed on Feb. 17, 2016.
This patent is subject to a terminal disclaimer.
Int. Cl. G06K 9/62 (2022.01); G06F 18/2411 (2023.01); G06K 9/00 (2022.01); G06T 7/13 (2017.01); G06V 10/44 (2022.01); G06V 40/20 (2022.01)
CPC G06F 18/2411 (2023.01) [G06T 7/13 (2017.01); G06V 10/44 (2022.01); G06V 40/28 (2022.01); G06T 2207/10028 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method of preparing sample hand positions for training of neural network systems, the method including:
obtaining ground truth simulated stereoscopic hand images;
extracting hand boundaries for the hand images and aligning the hand boundaries with hand centers included in the hand images;
generating translated, rotated and scaled variants of the hand boundaries and applying Gaussian jittering to at least some variants;
extracting hand regions from the at least some variants of the hand boundaries to which Gaussian jittering was applied;
computing ground truth pose vectors for the hand regions; and
storing the ground truth pose vectors in tangible machine readable memory as output labels for the ground truth simulated stereoscopic hand images.