Eye Tracking Goggles

The objective of the goggles is to track patients’ eyes for medical exemptions and analysis for the Oakville Centre for Vision clinic. This has been my main project for the two summers that I was working there.
Key Skills
  • Cross-disciplinary Product development

  • Hardware Prototyping:
  • Fusion 360 CAD
  • 3D Printing
  • Design for Assembly
  • Circuit Design, CSI, USB Communication

  • Software Prototyping:
  • OpenCV
  • Arduino C++/C Programming
  • Perspective-n-point Implementation
IR LEDs light up the eyes and illuminate the eyes for the highspeed IR cameras. We use those images to triangulate where the eyes are looking. All CV processing is done on a computer applicaiton as seen below.
Because we had a small team of 3 engineers, I worked cross-functionally to ensure we meet our hardware requirements and provide help to software such as fixing problems with calibration and overall tracking preformance in CV.
Overview Contributions
  • Full design of multiple hardware iteration of prototypes
  • improvement of cameras positioning, specificaiton, boardsize
  • Assembling of all prototypes, soldering, 3D printing, machining, etc
  • Identify and outreach to multiple consultants to cover holes in our project and team
  • Implementing of all calibration related solutions
  • Prototyping and algorithm design of other complementary projects
Hardware Features
  • USB 1080p@60fps world camera
  • Adjustable world camera POV from 40 to 180 degrees
  • 200x200p@200fps eye cameras
  • Adjustable camera mounts
  • Prescription lens holder
  • 3D printed camera mounts for modularity
  • Energy efficient, all powered from one computer USB
  • Cost effective of around $150 per unit to build
Software Personal Contributions
  • Eliminate calibration issues by implementing Perspective-n-point on QR codes using OpenCV
  • Help translate code from open source library Pupil Core
  • Writing on camera memories to identify camera and its position
I also got the chance to develop other complementary projects such as a ball catch sensor that automatically detects ball catch for sport related eye issues.
Ball Catch Device Features
  • Bluetooth wireless connection
  • Automatic calibration to detect catch of any sized ball
  • USB rechargable on board battery
  • Flexible 3D printed material construction
  • Utilization of Least Square Regression to smooth sensor noise
  • Buzzer to notify device status
  • Wearable design
  • Non-intrusive catching experience
  • Fully embedded Arduino system