Welcome to the official repository for our paper: “Detecting gaze shifts of moving observers in dynamic environments” submitted to Behavior Research Methods.
This repository provides: 1) The Python code to reproduce our benchmark comparing six popular gaze-shift detection methods on head-mounted eye-tracking data recorded outdoors, 2) The implementation of our proposed Ranking algorithm for robust, parameter-free gaze-shift detection, 3) Scripts for evaluation, comparison, and visualization.
The public dataset can be found in the following link:
https://unishare.nl/index.php/s/Ypgm3btwGs5wAYr
Before being able to run the code, write an script that creates a folder named image_2 in each particiapnt folder and transforms each world.mp4 video to a set of images from the frames of the video. The naming of the image frames should start from 000001.png. This folder will be only needed by ACE-DNV method.
The required libriaries and independencies are listed in the requirements.txt. You can use pip to install all of them together.
Set the address to the data folder in the config.py.
For comparing all the methods including the pre-trained machine-learning-based methods, run main.py. For optimizing the threshold-based methodsm run optimizeThreshold.py. To retrain the machine-learning-based methods, execute training.py.
For citations please use the following publication:
Citation will be provided after publication
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 955590.
