tcsantini / eyerec

Modular components for real-time mobile eye tracking.

EyeRec

Modular components for real-time mobile eye tracking.

Table of Contents

• Examples
• Available Methods
• To-be-Added Methods
• I Don't Really Want to Compile Stuff

Examples

1. Depending on your setup, you might need to install / specify (or both) OpenCV's location. E.g.:

# Install opencv on Ubuntu:
sudo apt install libopencv-dev
# Set opencv location
export OpenCV_DIR=<opencv-path>

2. Compiling and running the example application for pupil tracking

mkdir build
cd build
cmake ../
make install -j 8
./cpp/examples/track-pupil <algorithm-name> <path-to-video>

3. The above command will also install the library, headers, and CMake config files, supporting cmake find_package (e.g., see the examples). After setting eyerec_DIR or CMAKE_PREFIX_PATH accordingly, you can use this library in your project roughly so:

find_package(eyerec REQUIRED)
add_executable(your-app ...)
target_link_libraries(your-app eyerec::eyerec)

Available Methods

Pupil Detection

• PuRe

Pupil Tracking

• PuReST

To-be-added Methods

Calibration Methods

• (TODO) CalibMe

Gaze Estimation Methods

• (TODO) Grip

I Don't Really Want to Compile Stuff

What you are probably looking for then is EyeRecToo, a Windows application that provides:

• real-time slippage-robust mobile eye tracking
• multiple pupil detection and tracking methods
• multiple gaze estimation methods
• high-usability gaze calibration
• recording of eye-tracking videos and data
• fiduciary marker detection
• support for multiple head-mounted eye trackers, including the popular Pupil Core hardware.