2015

Samsung Health

Health algorithms based not on math, but on bioscience and physiology.

The Challenge

Samsung was an early adopter of PPG sensors for heart rate and stress in wearables and phones. However, the sensors were rendered nearly useless due to the slow response times of the algorithms, which took naive approaches to denoising and error detection — and that would mean the end for our nascent ambitions in health.

Outcomes

A new approach to noise detection, denoising, and multi-parameter computation.
3x faster stress readings.
6x faster heart rate.
Shipped on devices since 2015.
Certified to medical-grade standards and recently FDA-approved.

The Story

When optical heart rate (PPG) sensors first appeared in Samsung devices, the algorithms backing them were woefully slow and highly prone to noise. With such response times, the sensors were effectively useless in practical situations, with readings requiring 1 minute of stillness.

By approaching the noise detection issues from a physiological perspective, we not only made the sensor algorithms much more resilient, but we could detect erroneous readings from dynamic body motion and insufficient skin contact, leading to immensely faster results.

The results speak for themselves.

More…

Think Tank Team Site