US 12,437,526 B2
Methods and systems for automated display verification
Debabrata Pal, Bangalore (IN); Abhishek Alladi, Hyderabad (IN); and Luke Zeleznak, Phoenix, AZ (US)
Assigned to HONEYWELL INTERNATIONAL INC., Charlotte, NC (US)
Filed by HONEYWELL INTERNATIONAL INC., Charlotte, NC (US)
Filed on Aug. 5, 2022, as Appl. No. 17/817,860.
Claims priority of application No. 202211036116 (IN), filed on Jun. 23, 2022.
Prior Publication US 2023/0419647 A1, Dec. 28, 2023
Int. Cl. G06V 10/82 (2022.01); G06N 20/10 (2019.01); G06V 10/74 (2022.01); G06V 10/77 (2022.01); G06V 10/776 (2022.01)
CPC G06V 10/82 (2022.01) [G06N 20/10 (2019.01); G06V 10/761 (2022.01); G06V 10/7715 (2022.01); G06V 10/776 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A method of automated validation of an unverified image, the method comprising:
analyzing the unverified image to identify a first plurality of kernels associated with a first layer of a convolutional neural network for the unverified image;
analyzing a reference image to identify a second plurality of kernels associated with the first layer of the convolutional neural network for the reference image;
creating a shared set of kernels associated with the first layer based at least in part on the first plurality of kernels and the second plurality of kernels in accordance with one or more kernel shortlisting criteria;
obtaining a first feature map for the first layer of the unverified image using the first layer of the convolutional neural network and the shared set of kernels;
obtaining a second feature map for the first layer of the reference image using the first layer of the convolutional neural network and the shared set of kernels;
calculating a similarity score for the unverified image based at least in part on differences between the first feature map and the second feature map and a weighting factor associated with the first layer; and
automatically validating the unverified image based at least in part on the similarity score.