Inspiration
Right now, heart rate tracking can require expensive equipment that most people don’t have. We wanted to make an easy way to track your heart rate without having to buy this. Also, recently, there’s been lots of news of people having heart attacks while using stationary bikes. AutoHealth has the potential to prevent this.
What it does
AutoHealth is a webapp that uses machine learning (ML) to track your heart rate, for example, while you exercise. This makes it easy to see it without lots of setup, and in a hands-free way.
How we built it
The machine learning algorithm tracks the user’s face using a Haar cascading featurizer then determines the amount of blood in a person’s face from subtle color changes. Overtime, the heart rhythm is established and the heart rate can be interpreted and graphed.
Challenges we ran into
Computer vision is most commonly implemented in python, but javascript is a much easier language to build webapps in so we had to translate multiple computer vision algorithms to javascript for our use. This took multiple hours, but gave us a nice result. Also, our team members learned new frameworks for the first time which proved challenging, but rewarding with new skillsets and a nice final product.
Accomplishments that we're proud of
We are proud of bringing the concept of tracking heart rate with cameras to a reality. The project successfully demonstrates the use case and viability of this technology to help people in their everyday lives.
What we learned
We learned how to effectively apply ai and machine learning to a real product. This cutting edge technical knowledge is helpful in industry and for advancing research in the field.
What's next for AutoHealth
The next steps will be to integrate workout sites such as peloton to offer the on-demand workout classes. Additionally, we will be researching how to detect heart attacks.
Log in or sign up for Devpost to join the conversation.