US 12,230,015 B2
Fine-level face alignment through anchor and contour estimations
Jihua Huang, Sunnyvale, CA (US); and Amir Tamrakar, Princeton, NJ (US)
Assigned to SRI International, Menlo Park, CA (US)
Filed by SRI International, Menlo Park, CA (US)
Filed on Mar. 15, 2022, as Appl. No. 17/654,956.
Claims priority of provisional application 63/161,874, filed on Mar. 16, 2021.
Prior Publication US 2022/0301290 A1, Sep. 22, 2022
Int. Cl. G06V 10/776 (2022.01); G06T 7/64 (2017.01); G06T 7/73 (2017.01); G06V 10/75 (2022.01); G06V 10/774 (2022.01); G06V 40/16 (2022.01)
CPC G06V 10/776 (2022.01) [G06T 7/64 (2017.01); G06T 7/73 (2017.01); G06V 10/751 (2022.01); G06V 10/7747 (2022.01); G06V 40/161 (2022.01); G06V 40/172 (2022.01); G06T 2207/20081 (2013.01); G06T 2207/30201 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A system for performing facial alignment tasks, the system comprising:
an input device configured to receive a training image comprising a plurality of pixels, wherein the training image is labeled with a plurality of facial landmarks; and
a computation engine comprising processing circuitry for executing a machine learning system, wherein the machine learning system is configured to:
generate, from the training image, one or more facial contour heatmaps, wherein each of the one or more facial contour heatmaps depicts an estimate of a location of one or more facial contours within the training image; and
train a machine learning model to process the one or more facial contour heatmaps to predict the location of the one or more facial contours within the training image, wherein training the machine learning model comprises applying a loss function to minimize a distance between the predicted location of the one or more facial contours within the training image and corresponding contour data generated from facial landmarks of the plurality of facial landmarks with which the training image is labeled to improve an accuracy of the machine learning system in facial recognition.