US 12,268,188 B2
System and method for characterizing and monitoring health of an animal based on gait and postural movements
Paul Mundell, San Francisco, CA (US); Camden Olson, San Francisco, CA (US); Libby Rockaway, San Francisco, CA (US); and John Honchariw, San Francisco, CA (US)
Assigned to Companion Labs, Inc., San Francisco, CA (US)
Filed by Companion Labs, Inc., San Francisco, CA (US)
Filed on Aug. 11, 2022, as Appl. No. 17/886,378.
Application 17/886,378 is a continuation in part of application No. 17/538,242, filed on Nov. 30, 2021.
Application 17/538,242 is a continuation of application No. 16/418,596, filed on May 21, 2019, granted, now 11,576,348.
Claims priority of provisional application 63/231,800, filed on Aug. 11, 2021.
Claims priority of provisional application 62/674,575, filed on May 21, 2018.
Prior Publication US 2023/0083421 A1, Mar. 16, 2023
Int. Cl. A01K 15/02 (2006.01); A01K 29/00 (2006.01); G06T 7/73 (2017.01); G06V 40/10 (2022.01); G06V 40/20 (2022.01)
CPC A01K 15/021 (2013.01) [A01K 29/005 (2013.01); G06T 7/73 (2017.01); G06V 40/10 (2022.01); G06V 40/20 (2022.01); G06T 2207/10016 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for autonomously training an animal, comprising:
during a first training session for an animal:
accessing a video feed recorded by an optical sensor defining a field of view intersecting a working field and integrated into a training apparatus configured to dispense units of a primary reinforcer toward a range of locations within the working field via a dispenser integrated into the training apparatus; and
in the video feed, detecting the animal in the working field;
at a first time during the first training session, outputting a first cue associated with a first target pose;
in response to outputting the first cue:
detecting the animal in a first pose in the video feed; and
in response to the first pose falling within a threshold deviation of the first target pose, dispensing a first unit of the primary reinforcer to the animal via the dispenser;
extracting a first set of position data, representing positions of a set of anatomical features of the animal, from a first frame, in the video feed, depicting the animal in the first pose during the first training session;
deriving a first pose profile for the animal based on the first set of position data, the first pose profile representing posture of the animal in the first target pose during the first training session;
accessing a baseline pose profile derived from position data collected for the animal during an initial time period preceding the first training session;
characterizing a difference between the first pose profile and the baseline pose profile; and
in response to the difference exceeding a threshold difference:
interpreting a first abnormality in the first pose profile;
predicting a first causal pathway for the first abnormality based on the difference; and
selecting a first mitigation protocol for the animal based on the first causal pathway.