Edge Impulse Announces New PPG/ECG Heart Rate Algorithm Needing Just a Fraction of the RAM
New algorithm can squeeze insights out of even the noisiest of sensor data, using one-sixteenth the memory of the competition.
Gareth HalfacreeFollow
6 hours ago •
Health & Medical Devices /
Wearables /
Sensors /
Machine Learning & AI
On-device machine learning (ML) and artificial intelligence (AI) specialist Edge Impulse has announced the launch of what it claims is the "smallest, most precise" heart rate measurement algorithm around — requiring one-sixteenth the memory of the competition.
“Our new algorithm generates clean HR and HRV [Heart Rate and Heart Rate Variability] values from a PPG [Photoplethysmogram] sensor via our augmentation of standard processing techniques. Our enhancements mitigate significant noise typically associated with using data from wearables, such as on a finger, outperforming known algorithms in MAE [Mean Absolute Error], variance, and resource usage," claims Edge Impulse's Alex Elium, lead developer on the new algorithm.
"This significantly reduces the R&D investment to build custom algorithms that would take years to refine for use in the field, whether for clinical trials or consumer wearable devices."
Edge Impulse's new heart rate algorithm offers greater accuracy than the competition, while using a fraction of the RAM. (
: Edge Impulse)
The algorithm primarily targets light-based photoplethysmogram sensors, as found in smartwatches and fitness bands, though can also be applied to electrocardiogram (ECG) sensors, the company says. It's capable of recognising increased and decreased activity, falls, sleep, stress levels, and atrial fibrillation — and can do so using just one-sixteenth the RAM of its closest competition.
The HR and HRV algorithm is part of a suite of healthcare-related algorithms the company has launched, including one targeting electroencephalogram (EEG) brain activity data, another looking at body temperature, and yet another for tracking movement and posture. To support these, Edge Impulse has also announced a "research data lake" for clinical data, combined with data dashboards for real-time monitoring.
The algorithm can be used to clean up data and extract a range of information, including stress levels. (
: Edge Impulse)
Edge Impulse has named a series of companies already using its technology, including wearables makers Oura, Nowatch, and SlateSafety, Hyfe, and Know Labs, the latter of which is bringing the first FDA-approved non-invasive blood glucose monitoring system to market.
New algorithm can squeeze insights out of even the noisiest of sensor data, using one-sixteenth the memory of the competition.
www.hackster.io