CalorieReceipt

Inspiration:

Being health conscious is tough. As someone who tries to be healthy knows, keeping track of all your day to day fitness and diet patterns can be discouraging for someone trying to improve their life. That is why we focus on simplicity. Our website requires two steps: upload and submit.

What You Leaned:

We learned how to delegate task, and communicate as a team in order to get things done. It addition, we learned the basics of operating a raspberry pi, the command lines it uses, and how to run optical character recognition.

Built:

Our website is built on flask and hosted on a raspberry pi. The input is an image file that is converted to text using tesseract. We then analyze the way the receipts are formatted and parsed through each line so that we only got the information we wanted to get the item's name. Once we got the item's name we compared it to the database we created by scraping the chain's health information page. It then returns the name of the item along with it's caloric information.

Challenges:

The idea we came up with dealt with parts of coding that none of us had experience in, especially in optical character recognition and web hosting . Most of our time was spent reading documentation on languages and programs and trying to figure out how to use them.

What is Next?

So you might be wondering why a program like this is not a mobile app... we initially wanted this to be an iOS app, but after realizing that it would be too much to learn from scratch on top of learning the other core components of the build, we decided to forgo it and do a website instead in the interest of time. Thus the next step for CalorieReceipt is to transition it over to the mobile platform, alongside scaling the program to support a greater number of restaurants.