US 11,676,292 B2
Machine learning inference on gravity aligned imagery
Chloe LeGendre, Culver City, CA (US); Ranjith Kagathi Ananda, Mountain View, CA (US); Ran Tao, San Ramon, CA (US); and Wim Meeussen, Redwood City, CA (US)
Assigned to GOOGLE LLC, Mountain View, CA (US)
Filed by GOOGLE LLC, Mountain View, CA (US)
Filed on Jun. 22, 2021, as Appl. No. 17/304,509.
Application 17/304,509 is a continuation of application No. 16/643,213, granted, now 11,069,075, previously published as PCT/US2019/057218, filed on Oct. 21, 2019.
Prior Publication US 2021/0312646 A1, Oct. 7, 2021
Int. Cl. G06T 7/37 (2017.01); G06T 7/73 (2017.01); G06T 11/00 (2006.01)
CPC G06T 7/37 (2017.01) [G06T 7/73 (2017.01); G06T 11/00 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30201 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A wearable computing device comprising:
an image capture device;
at least one sensor;
at least one processor; and
memory storing instructions that, when executed by the at least one processor, cause the processor to:
obtain, by the image capture device, a first image feed including a plurality of objects tracked by the at least one sensor;
identify a current image as one of the plurality of images of the first image feed;
detect, using the at least one sensor, a device orientation of the wearable computing device;
determine an image orientation of the current image in relation to the device orientation;
determine, based on the device orientation and the image orientation, a rotation angle in which to rotate the current image to an upright orientation;
rotate the current image to the rotation angle to generate a second image;
generate a gravity-aligned version of the plurality of images of the first image feed based on the second image; and
generate, using a neural network, based on the gravity-aligned version of the plurality of images of the first image feed, Augmented Reality (AR) content for rendering with at least one of the plurality of objects of the first image feed.