Calorify

Inspiration

We wanted to build a health hack. What more can keep us healthy more than the food we eat. We wanted to explore various ways of understanding the food we eat.

What it does

Evaluates the calories and nutrients in your food. This information is just a few clicks away. Through our application, one can take a photo and get information about their diet.

How we built it

We built a react based web application and integrated it with node js backend. Node js will parse the image to a machine learning model provided by clarifai. We trained this model to detect various food items. This model identifies the food and returns its name. We will then pass the food images to a volume estimator module developed in python. Once the volume is estimated, the information about the food is obtained by calling nutritionx apis.

Challenges we ran into

Too much to do and not enough time. Working on react native mobile application restrained us in developing image parser to the backend. We tried using google vision api for the food identification but the model required beta version to result in a good estimate. Drawing contour lines for the volume estimation was very hard.