US 12,405,812 B2
Computer vision template matching system and method
Vibhu Rawat, Bangalore (IN); Abhilash Jain, Bangalore (IN); and Aashish Jain, Bangalore (IN)
Assigned to Whatfix Private Limited, Karnataka (IN)
Filed by Whatfix Private Limited, Bangalore (IN)
Filed on Oct. 24, 2023, as Appl. No. 18/493,549.
Prior Publication US 2025/0131622 A1, Apr. 24, 2025
Int. Cl. G06F 9/451 (2018.01); G06T 3/40 (2006.01); G06T 5/20 (2006.01); G06T 5/70 (2024.01); G06T 5/73 (2024.01); G06T 11/60 (2006.01); G06V 10/75 (2022.01); G06V 30/414 (2022.01)
CPC G06F 9/453 (2018.02) [G06T 3/40 (2013.01); G06T 5/20 (2013.01); G06T 5/70 (2024.01); G06T 5/73 (2024.01); G06T 11/60 (2013.01); G06V 10/751 (2022.01); G06V 30/414 (2022.01); G06T 2200/24 (2013.01); G06T 2207/20016 (2013.01); G06T 2207/20021 (2013.01); G06T 2210/32 (2013.01)] 28 Claims
OG exemplary drawing
 
1. A method of performing optimized template matching within a digital adoption platform used in conjunction with an underlying computer application, the method comprising the steps of:
capturing an original image and template from a user interface of the underlying application;
converting the original image and the original template to grayscale;
detecting a background type of the converted image and of the converted template, and inverting the background type of either the converted template or the converted image if there is a mismatch to create a matched image and template;
applying a Gaussian filter to the matched image and to the matched template to reduce any noise and any blurring in the matched image and in the matched template;
dividing a scale of the filtered image by a scale of the filtered template to determine a starting scale;
reducing a resolution of at least the filtered image;
applying a template matching algorithm to the reduced resolution image and the reduced resolution or filtered template using the starting scale to locate the original template in the original image; and
remapping coordinates of the original template to reflect its original location on the original image.