US 12,432,283 B2
Methods and apparatus for online test taking
Eric Roger Dawicki, Fairhaven, MA (US); Akkhilaysh Kushal Shetty, New Bedford, MA (US); and Angela Michele Dawicki, Fairhaven, MA (US)
Assigned to Northeast Maritime Institute, Inc., Fairhaven, MA (US)
Filed by Northeast Maritime Institute, Inc., Fairhaven, MA (US)
Filed on Nov. 2, 2021, as Appl. No. 17/516,991.
Claims priority of provisional application 63/109,217, filed on Nov. 3, 2020.
Prior Publication US 2022/0141298 A1, May 5, 2022
Int. Cl. G06T 7/70 (2017.01); G06N 20/00 (2019.01); G06V 20/52 (2022.01); G06V 40/20 (2022.01); G09B 5/08 (2006.01); G09B 7/00 (2006.01); H04L 67/50 (2022.01); H04L 67/54 (2022.01)
CPC H04L 67/535 (2022.05) [G06N 20/00 (2019.01); G06T 7/70 (2017.01); G06V 20/52 (2022.01); G06V 40/20 (2022.01); G09B 5/08 (2013.01); G09B 7/00 (2013.01); H04L 67/54 (2022.05)] 20 Claims
OG exemplary drawing
 
1. A system for administering an online examination to a user, the system comprising:
a processor configured to:
perform, based on first image data of an environment of the user received from a camera, a pre-examination scan of the environment of the user;
responsive to the pre-examination scan indicating the environment is appropriate for test taking, initiate the online examination by generating a user interface for display on an examination screen;
determine improper examination behavior using a multi-layered analysis to compound multiple points of reference, the determination comprising:
inputting second image data of the user while using the user interface into a trained machine learning model;
receiving an estimate of, based on the second image data, one or more poses of the user using a trained machine learning model, wherein the one or more poses include a head pose of the user and a body pose, and wherein the body pose includes an output of the trained machine learning model that indicates a position of one or more body joints of the user; and
determining, based on the one or more poses of the user estimated by the trained machine learning model, improper examination behavior by comparing the position of the one or more body joints of the user with a set threshold boundary for the position of the one or more body joints, and when the position of the one or more body joints of the user exceeds the set threshold boundary, determining that the user's behavior is improper; and
based on the improper examination behavior, generate a warning or terminate the online examination.