US 12,293,563 B2
Automated data annotation for computer vision applications
Karthikeyan Shanmuga Vadivel, San Jose, CA (US); Omar Oreifej, Costa Mesa, CA (US); and Patrick A. Worfolk, San Jose, CA (US)
Assigned to Synaptics Incorporated, San Jose, CA (US)
Filed by Synaptics Incorporated, San Jose, CA (US)
Filed on Jun. 1, 2022, as Appl. No. 17/829,844.
Prior Publication US 2023/0394786 A1, Dec. 7, 2023
Int. Cl. G06K 9/00 (2022.01); G06N 3/08 (2023.01); G06V 10/48 (2022.01); G06V 20/70 (2022.01)
CPC G06V 10/48 (2022.01) [G06N 3/08 (2013.01); G06V 20/70 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A method of training a machine learning model, comprising:
receiving a first input image depicting an object of interest and a planar extension object attached thereto, the planar extension object having a planar surface that is coplanar with a surface of the object of interest;
receiving a reference image depicting the object of interest and one or more annotations associated with the object of interest;
mapping a plurality of first points in the reference image to a respective plurality of second points in the first input image so that the one or more annotations in the reference image are projected onto the object of interest in the first input image, wherein one or more first points of the plurality of first points and one or more second points of the plurality of second points coincide with the planar surface; and
training the machine learning model to produce inferences from images depicting the object of interest based at least in part on the mapping of the plurality of first points to the plurality of second points.