US 11,941,536 B2
Entity learning recognition
Gaurav Goswami, New Deli (IN); Sharathchandra Umapathirao Pankanti, Yorktown Heights, NY (US); and Nalini K. Ratha, Yorktown Heights, NY (US)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Sep. 23, 2021, as Appl. No. 17/482,613.
Application 17/243,786 is a division of application No. 15/472,837, filed on Mar. 29, 2017, granted, now 11,093,796, issued on Aug. 17, 2021.
Application 17/482,613 is a continuation of application No. 17/243,786, filed on Apr. 29, 2021.
Prior Publication US 2022/0012543 A1, Jan. 13, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. G06N 5/022 (2023.01); G06F 18/21 (2023.01); G06F 18/214 (2023.01); G06N 20/00 (2019.01); G06V 10/774 (2022.01); G06V 10/776 (2022.01)
CPC G06N 5/022 (2013.01) [G06F 18/2148 (2023.01); G06F 18/217 (2023.01); G06N 20/00 (2019.01); G06V 10/7747 (2022.01); G06V 10/776 (2022.01)] 2 Claims
OG exemplary drawing
 
1. A computer-implemented entity learning recognition training method, comprising:
obtaining an original image of a face from an augmented database;
selecting a structural element from a plurality of structural elements, wherein the
structural element comprises a shape associated with the face in the image;
generating an augmented image of the face by combining the structural element with the
original image of the face and adding the augmented image of the face to the augmented
database; and training a model using augmented face images in the augmented database, wherein the training is based on a visual element portion of the augmented face images in the augmented database with added noise.