Chord Connect Mobile App

Computer Vision work with openCV for Android and iOS to reimagine the QR Code.

CLIENT: Cloud Engage

TEAM: Bhaskar Athma, Ranjit Pandit, Scott Rozic, Tim Drake, Adam Parish

TAGS: computer vision,opencv,app

Chord Connect App Screens

Thinking Annularly

The traditional QR code is a bit clunky. CloudEngage wanted to reimagine their own version of a QR code that was more aesthetically pleasing, simplified, and could be used to connect two people easily to share contact details, photos, and information. I was tasked with creating the logic that would recognize and translate that code from binary data into an 6 digit letter sequence matching the user's profile.

The system implemented uses Hough Circle Detection to find the user's avatar at the center of the code and, based on its size, creates 3 outer rings in which to scan for the data at 48 points along those circles. If a region is dark, it is a 1, and if white, a 0. The outer circle contains a start indicator which is used for orientation and start sequence.

To hone in on the best detection settings, I created an on-screen GUI that could be used by testers to tweak parameters such as radius, circle accumulation, light and dark camera settings, and more.

The refined scanning logic could detect codes on small devices such as an Apple Watch, or larger devices such as a desktop, from approx. 3in - 2ft reliably, and worked on Android and iOS. The code was generated using D3.js