US 11,842,254 B2
Object recognition based on hierarchical domain-based models
Luis M. Tomotaki, Plano, TX (US); William F. Copeland, Garland, TX (US); Gina L. Otts, Flower Mound, TX (US); Saravanan Mallesan, Fairfax, VA (US); and Natasha Avinash Shah, Irving, TX (US)
Assigned to Verizon Patent and Licensing Inc., Basking Ridge, NJ (US)
Filed by Verizon Patent and Licensing Inc., Basking Ridge, NJ (US)
Filed on Nov. 5, 2020, as Appl. No. 17/089,819.
Application 17/089,819 is a continuation of application No. 15/632,652, filed on Jun. 26, 2017, granted, now 10,846,555.
Prior Publication US 2021/0056340 A1, Feb. 25, 2021
Int. Cl. G06N 20/00 (2019.01); G06T 7/00 (2017.01); G06N 3/08 (2023.01); G06N 20/20 (2019.01); G06V 10/20 (2022.01); G06V 20/40 (2022.01); G06V 40/10 (2022.01); G06F 18/21 (2023.01); G06F 18/231 (2023.01); G06V 10/762 (2022.01); G06V 10/776 (2022.01); G06N 3/045 (2023.01)
CPC G06N 20/00 (2019.01) [G06F 18/217 (2023.01); G06F 18/231 (2023.01); G06N 3/08 (2013.01); G06N 20/20 (2019.01); G06T 7/0002 (2013.01); G06V 10/255 (2022.01); G06V 10/7625 (2022.01); G06V 10/776 (2022.01); G06V 20/42 (2022.01); G06V 40/103 (2022.01); G06N 3/045 (2023.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
receiving, by a parent inference model of a network device, data indicating a sub-classification of a sample object and a sample image that includes the sample object, wherein the network device further includes a child inference model, and the parent inference model and the child inference model each performs object recognition;
generating, by the parent inference model, a bounding box that indicates a region of interest with respect to the sample object;
identifying, by the parent inference model, a first general classification of the sample object included in the sample image;
generating, by the parent inference model, sample data indicating the first general classification of the sample object;
replacing the sample data with the data indicating the sub-classification; and
training the child inference model using the sample image, the bounding box, and the data indicating the sub-classification.