Inspiration

I was on Tinder and I found a match. That same day I ate 3 burgers in one sitting. It was such an experience that I felt I had to merge these two realms to one ultimate existence.

What it does

If you can't think of what you eat, it shows you appetizing pictures of food (without brand names), and you can swipe left for reject or right for approval.

How I built it

Built a website using HTML5 and JQuery, hosted on an EC2 instance in Amazon Web Service, as well as a Flask service that pulls image and meta data from Yelp and caches classified images on an S3 Bucket for 23 hours to conform to Yelp's Terms of Service, and to redue the number of classifications we need to do. Image classification is done through Amazons "Rekognition" service. We upload images from Yelp and only accept images that are classified as food, and do not contain people (It reports back a set of tag:confidence key pairs that are how likely it thinks the image contains a certain object). We also used the Route53 service from Amazon to ensure our DNS records propagated quickly from the purchase of our domain.

Challenges I ran into

The internet connection in the gymnasium is not good. It caused a lot of issues with loading times, and connections to remote services

Accomplishments that I'm proud of

Learning a vast amount of knowledge about AWS services, going from essentially no knowledge to using 4 separate services (almost 5)

What I learned

Managing EC2 Using S3 Buckets Route53 DNS Flask

Share this project:
×

Updates