Edge Impulse logo
Blog Post

Using Machine Learning for Health and Wellness Monitoring in Consumer Wearable Devices

Machine Learning, Embedded DevicesThis blog post explores Edge Impulse's scalable approach to health machine learning engineering.

Jed Huang

August 3, 2022

Today’s consumer smart devices provide amazing insight into our health and wellness. Clinicians are even using wearables to collect data for clinical trials, measuring functions like pulse rate, temperature level, or even blood oxygen saturation. This data is being put to effective use, but with the help of machine learning and sensor fusion capabilities from platforms like Edge Impulse, you can help build novel, specific solutions that help all live even healthier lives. 

Read the Our Scalable Approach to Health Machine Learning Engineering sectionOur Scalable Approach to Health Machine Learning Engineering

Edge Impulse already works with companies in the health device space like Oura, SlateSafety, and Knowlabs on their full data engineering and machine learning algorithm development process. We offer a platform approach for this process to ensure that data is labeled properly, and that algorithms are built against data being collected from gold standard tests to ensure accuracy. Furthermore, algorithms can be tested and improved over time, and even personalized. This is especially critical for health-related algorithms to detect conditions such as fatigue or stress levels where variations between physiology may exist.

Through our enterprise platform, APIs, and solutions offering, Edge Impulse helps wearables and health device companies bring new algorithms to market faster; build robust, accurate algorithms on par with gold standard medical-grade tests; improve, personalize, and update algorithms over time; integrate sensor data streaming; and automate machine learning pipeline development process with new data input.  

Consumer wearable devices can help:

  • People self-manage their disease or conditions without providing specific treatment suggestions.
  • Patients document, show, or communicate potential medical conditions to health care providers.
  • Patients, clinicians, or providers to interact with electronic health records.
  • Automate data tasks related to sensors for health care providers.
  • Ensure clinical-grade data collection procedures for sensors, plus machine learning model development and maintenance.

According to Voler Systems, some of the common physiological function data being collected and measured by consumer wearable devices from Google, Apple, and Oura include:

  • Body temperature
  • Heart rate (pulse) rhythm and volume
  • Blood oxygen
  • Blood pressure (diastolic/systolic)
  • Blood sugar (glucose)
  • EKG / ECG (electrocardiogram) — measures the electrical activity of the heart muscle, providing information on the heart’s response to physical exertion
  • EEG (electroencephalography) — measures electrical activity in the brain
  • EMG (electromyography) — measures electrical activity of muscles
  • Accelerometer measures the movement of the body and muscle activity
  • Respiration — count, rhythm or regularity, character/type

With Edge Impulse, a provider can build meaningful analysis from sensors monitoring any of these functions, and more importantly, any combination of them, to help generate new levels of health-monitoring precision. We also have specific solutions for the health and wellness space, allowing data portals for researchers to upload clinical studies data.

Please feel free to reach out or book an exploratory call to learn more about how we can help your team build robust, gold standard health algorithms. 

Learn more about this on our upcoming lunch and learning webinar — register now


Are you interested in bringing machine learning intelligence to your devices? We're happy to help.

Subscribe to our newsletter