$can

PayPal Hack ATL 2018

This application uses Computer Vision, Augmented Reality, and Machine Learning to automatically detect bills and coins and count up the total amount of money you can see.

Team

Application built by:

Omar Sharifali
Jason Tu
Georgianna (Blue) Lin
Andrew Chang
Sarah Li

Slides

Built With

arcore
java
kotlin
rekognition

Submitted to

PayPal Hack-ATL
- Winner 3rd Place
- Winner Most Creative
- Winner Biggest Wow Factor

Created by

I spent the majority of the time working on the bill recognition in Android as well as providing the training data for the model. This was done with Amazon's Rekognition APIs.

Omar S
I mostly worked on seeding the backend portion of out application so that we would have some baseline of paper bill faces to compare to and build on. This process was mostly done through scripting in Python.

Andrew Chang
Worked on the Android app (Kotlin!). Architected the back-end and did devops (provisioning accounts via AWS IAM for S3 and Rekognition access).
Android app used ARCore and Sceneform.

Jason Tu
Developer | Gamer | Student
Worked on Android app. Designed and implemented app UI/UX. Helped work on seeding backend and python scripts.

Georgianna Lin
Sarah Li