CaloReal

Inspiration

During the COVID-19 lockdown, many people found themselves stuck indoors at home, as it was often unnecessary and even dangerous to go outside. Additionally, due to massive layoffs, a bad economy, and a deadly virus, anxiety has increased over the last several months. As a result, stress eating and lack of exercise has been increasing, which can lead to health concerns in much of the general population. CaloReal was designed to track and count your calories, with the simple snap of a picture.

Our original idea was to train an AI model to recognize different food objects within a picture, and then use depth mapping to calculate an approximate amount of food that was being eaten.

What it does

Our original idea was to train an AI model to recognize different food objects within a picture, and then use depth mapping to calculate an approximate amount of food that was being eaten. However, because we ran out of time, the current website only allows you to enter the food name, and gives the caloric information per serving.

How we built it

The website was created using google firebase, with HTML and javascript. The AI model was trained using Tensorflow, an online image library, and a pre-trained model.

Challenges we ran into

We had trouble linking the Tensorflow model with the website. Since we had little prior experience working with these tools, it took a long time to understand what was going on.

Accomplishments that we're proud of

This was the first hackathon for 3 of our team members, and we were able to get a new website up and running fairly quickly with Google Firebase. It was also a brand new experience for all of our team working with TensorFlow's machine learning libraries, and was a great first experience overall.

What's next for CaloReal

We intend to finish the basic project within the next few days, and in the future implement the app on Android, to take advantage of the depth mapping capabilities of many Android device cameras, in order to calculate specific portion sizes.