FOXIE: A high-performance eye-tracking system

Project details

FOXIE is a compact, affordable, and high-performance eye-tracking system designed for researchers who need both precision and flexibility. Capable of processing up to 700 frames per second on embedded hardware, FOXIE delivers high-quality eye movement data. Its plug-and-play USB interface, cross-platform compatibility (Windows, Mac, Linux), and portable design make it ideal for both lab-based and field experiments.

Developed at Purple Gaze by a team led by Dr. Dmitry Mikushin, this product is aimed to democratize access to high-speed eye-tracking by offering a fraction of the cost compared to traditional systems.


Pupil-Tracking Algorithm
A demo showing internal workings of our pupil-tracking algorithm

Key Features

  • High Sampling Rate: Up to 700 Hz
  • Low Latency: Total system latency of 1.4 ms
  • High Accuracy: 0.5° visual angle
  • Operating Distance: 30–60 cm
  • Cross-Platform Support: Compatible with Windows, macOS, and Linux
  • Plug-and-Play: Simple USB connection
  • Portable Design: Small, lightweight, and fits in a compact case

Responsibilities

  • Developed algorithms to improve accuracy of the core pupil-tracking algorithms
  • Explored and implemented various parallel programming techniques to accelerate the core pupil-tracking algorithms

Tech Stack and Skills

Programming frameworks and tools:
  • C/C++, OpenMP, Intel TBB, ARM NEON, Nvidia CUDA, Nvidia Nsight Systems, Nvidia Nsight Compute, Git, CMake, Linux
Skills:
  • Computer Vision, Digital Image Processing, Algorithms, Parallel Programming, GPU Programming, Performance Profiling
C
C
C++
C++
OpenMP
OpenMP
CUDA
CUDA
Nsight Systems
Nsight Systems
Nsight Compute
Nsight Compute
CMake
CMake
Git
Git
Linux
Linux
Jetson
Nvidia Jetson
Raspberry Pi
Raspberry Pi