US 12,383,205 B2
Method for estimating physiological events from physiological signals, a non-transitory computer-readable medium, and, an apparatus
Vanessa Brischi Olivatto Fioravanti, Campinas (BR); Paula Gabrielly Rodrigues, Campinas (BR); Rafael Goncalves De Lima, Campinas (BR); and Otavio Augusto Bizetto Penatti, Campinas (BR)
Assigned to SAMSUNG ELETRÔNICA DA AMAZÔNIA LTDA., Campinas (BR)
Filed by SAMSUNG ELETRÔNICA DA AMAZÔNIA LTDA., São Paulo (BR)
Filed on Mar. 2, 2022, as Appl. No. 17/684,818.
Claims priority of application No. 10 2021 026473-0 (BR), filed on Dec. 27, 2021.
Prior Publication US 2023/0200741 A1, Jun. 29, 2023
Int. Cl. A61B 5/00 (2006.01); A61B 5/0205 (2006.01); A61B 5/024 (2006.01); A61B 5/0285 (2006.01); A61B 5/352 (2021.01); A61B 5/369 (2021.01)
CPC A61B 5/7267 (2013.01) [A61B 5/0205 (2013.01); A61B 5/02416 (2013.01); A61B 5/02438 (2013.01); A61B 5/0285 (2013.01); A61B 5/352 (2021.01); A61B 5/369 (2021.01); A61B 5/681 (2013.01); A61B 5/7207 (2013.01); A61B 5/7221 (2013.01); A61B 5/7225 (2013.01)] 25 Claims
OG exemplary drawing
 
1. A method for estimating physiological events from physiological signals received from one or more devices, the method comprising:
detecting candidate peaks or valleys on the one or more physiological signals,
fine-tuning positions of the detected candidate peaks or valleys using a first machine learning (ML) algorithm trained in a supervised manner to minimize errors relative to a reference signal;
extracting features from the fine-tuned candidate peaks or valleys, and
classifying, using the extracted features and a second ML algorithm trained in a supervised manner, the fine-tuned candidate peaks or valleys to identify peaks or valleys of interest; and
estimating one or more physiological events using the identified peaks or valleys of interest.