richardandersson/eyemovementdetectorevaluation

Code and data for evaluating eye movement detection algorithms. Material used in the paper: Andersson, R., Larsson, L., Holmqvist, K., Stridh, M. & Nyström, M. (2016). One algorithm to rule them all? : An evaluation and discussion of ten eye movement event-detection algorithms. Behavior Research Methods, 1-22. The Psychonomic Society.

eyemovementdetectorevaluation's Issues

‘HelpFunctions ’ in 'matlab_annotator' does not exist?

Hi sir,
Thank your for your sharing for those code, but I couldn't run it as expect.
When I run the 'EyeMovementDetectorEvaluation/matlab_annotator/EventTag.m' file, it shows that EyeMovementDetectorEvaluation/matlab_annotator/HelpFunctions don't exist.
And, 'EyeMovementDetectorEvaluation/matlab_analysis_code/mainDetection.m' could not run as well, it's seems that some function.m lossed.( I can't find Algorithm function or such as getEtFileNames function etc.)

Can't wait for your replying. THANKS A LOT!

Q: Adjustment to tiny labeling mistake (ref Zemblys, 2018)?

Zemblys et al., 2018, report a minor labeling mistake in one image file here (or a pay-wall free version here) in Appendix 2:

We found an obvious labeling mistake in the one of the validation trials (file UH29_img_Europe_labelled_MN. We fixed this error by reassigning 75 samples, [3197,3272) (zero-based index), from the saccade to the fixation class.

I checked the data file in question and it appears to still contain the erroneous saccade labels. Just to reconfirm: this labeling error has not been fixed in the data file in this repository, correct?

If you wish, I can PR a fixed file, the issue at hand is intended to just reconfirm my assumption.

Thanks in advance!

Missing data file?

I am trying to implement a validation script in the spirit of Andersson et al (2016) using the data published here (BIG THANKS for that!). The paper is talking about 34 files that are the basis of the validation. However, the repo contains 96 data files:

% find annotated_data/ -type f -name '*.mat' |wc -l
96

I assume that the 34 files used are the subset that has been labeled by both coders with MN defining the subset:

% find annotated_data/ -type f -name '*MN.mat' |wc -l
34

However, only 33 of those 34 have a matching file for coder RA.

% find annotated_data/ -type f -name '*MN.mat' | sed -e 's/MN/RA/'| xargs  file
<snip>
annotated_data/dots/UH33_trial1_labelled_RA.mat:               cannot open `annotated_data/dots/UH33_trial1_labelled_RA.mat' (No such file or directory)
<snip>

Is this file needed, or is my interpretation wrong?

Thanks in advance!