US 11,937,943 B2
Detection of physical abuse or neglect using data from ear-wearable devices
Matan Sivan, Tel Aviv (IL); Nitzan Bornstein, Tel Aviv (IL); Justin Burwinkel, Eden Prairie, MN (US); Roy Talman, Herzliya (IL); and Neta Aviel, Tel-Mond (IL)
Assigned to Starkey Laboratories, Inc., Eden Prairie, MN (US)
Filed by Starkey Laboratories, Inc., Eden Prairie, MN (US)
Filed on Mar. 20, 2023, as Appl. No. 18/186,568.
Application 18/186,568 is a continuation of application No. 16/777,525, filed on Jan. 30, 2020, granted, now 11,607,170.
Claims priority of provisional application 62/800,227, filed on Feb. 1, 2019.
Prior Publication US 2023/0225675 A1, Jul. 20, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. A61B 5/00 (2006.01); A61B 5/0205 (2006.01); A61B 5/16 (2006.01); A61B 5/369 (2021.01); G06N 3/08 (2023.01); G06N 20/00 (2019.01); H04R 3/00 (2006.01); A61B 5/024 (2006.01)
CPC A61B 5/6815 (2013.01) [A61B 5/0002 (2013.01); A61B 5/0024 (2013.01); A61B 5/0205 (2013.01); A61B 5/165 (2013.01); A61B 5/486 (2013.01); A61B 5/742 (2013.01); G06N 3/08 (2013.01); G06N 20/00 (2019.01); H04R 3/005 (2013.01); A61B 5/02438 (2013.01); A61B 5/369 (2021.01); A61B 2562/0219 (2013.01); A61B 2562/0223 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A method comprising:
obtaining, by one or more processing circuits, a first set of features characterizing a segment of inertial measurement unit (IMU) data generated by an IMU of an ear-wearable device;
obtaining, by the one or more processing circuits, a second set of features, wherein the second set of features characterizes a segment of audio data generated by the ear-wearable device or another ear-wearable device concurrently worn by a user of the ear-wearable device;
applying, by the one or more processing circuits, a machine learning model (MLM) that takes the first set of features and the second set of features as input;
determining, by the one or more processing circuits, based on output values produced by the MLM, whether the user of the ear-wearable device has potentially been subject to improper care; and
performing, by the one or more processing circuits, an action in response to determining that the user of the ear-wearable device has potentially been subject to improper care, wherein:
the MLM is an artificial neural network that includes a plurality of output neurons, the plurality of output neurons including a separate output neuron for each type of improper care in a plurality of types of improper care,
applying the MLM comprises applying the artificial neural network, wherein applying the artificial neural network comprises generating an output value for each of the output neurons of the artificial neural network, and
determining whether the user of the ear-wearable device has potentially been subject to improper care comprises determining, by the one or more processing circuits, based on the output values produced by the output neurons of the artificial neural network, whether the user of the ear-wearable device has potentially been subject to a type of improper care in the plurality of the types of improper care.