This is a R package which contains analysis scripts for eye tracking data acquired using ARETT and the Microsoft HoloLens 2.
As the gaze data generated by ARETT consists of three dimensional gaze information in an AR environment it differs from traditional eye trackers and is not supported by other R packages. This package implements the I-DT, I-VT and I-AOI fixation algorithms for this new data paradigm.
The package implements three offline fixation identification algorithms together with two pre- and two post-processing functions. Details about the implemented algorithms can be found in the included documentation and the accompanying publication.
Pre-processing:
- Gap fill
- Noise reduction
Fixation identification:
- Identification by velocity threshold (I-VT)
- Identification by dispersion threshold (I-DT)
- Identification by area of interest (I-AOI)
Post-processing:
- Merge adjacent fixations
- Discard short fixations
Clone the git project to a local folder (or download all files), open a R session in the folder and run
devtools::install()
All functions of the package are documented using roxygen2, available in the man folder and accessible after installation.
A snapshot of the documentation is also available in the projects wiki, however this snapshot might not be up to date.
@article{Kapp.2021,
title={ARETT: Augmented Reality Eye Tracking Toolkit for Head Mounted Displays},
author={Kapp, Sebastian and Barz, Michael and Mukhametov, Sergey and Sonntag, Daniel and Kuhn, Jochen},
journal={Sensors},
volume={21},
number={6},
pages={2234},
ISSN={1424-8220},
url={http://dx.doi.org/10.3390/s21062234},
DOI={10.3390/s21062234},
publisher={MDPI AG},
month={Mar},
year={2021}
}