Inspiration
We have been inspired by an EPFL course which provided a dataset of recipes that we had to clean.
What it does
Our IPhone app allow the user to take a picture of some ingredients, and then suggest some recipes to the user of the app given the aliments our server has detected.
How we built it
We decided to split our project in two main tasks:
- one part of the project was to set up the server and make a minimal app that can send some simple HTTP request to our server.
- another part was to detect objects in the image we took with the phone, and then send them to IBM cloud to identify each ingredient separately.
Challenges we ran into
Our biggest challenge was to get IBM cloud to recognize the image on our server. To solve this issue, we had to send our image to another server first, and from this server to send our images to IBM cloud. We also had some annoying issues such as figure out the proper encryption and decryption scheme for communication between IOS client and the server.
Accomplishments that we're proud of
For our first Hackathon, conceiving an app in restricted time corresponding to the original idea we had was the biggest accomplishment. We wanted to create something beyond our skills level and having a fully functional product is really an achievement that we are proud of.
What we learned
We had to learn how to create a server in python, use tensorflow library and communicate with the IBM cloud.
What's next for PocheCrew
We really enjoyed our first Hackthon and will surely do another one soon.
Built With
- ibm-watson
- python
- swift
- tensorflow
Log in or sign up for Devpost to join the conversation.