Perceptual User Interfaces Logo
University of Stuttgart Logo

Annotations as described in Müller et al. (2020)

The gaze annotations were collected for videos from the youtube channels "TheSpaDr" and "WisdomFromNorth". The naming scheme of the files is <youtube_id>.csv. Each file contains three columns: - index indicating the annotated frame
- 'gaze_left': gaze annotations for the left person
- 'gaze_right': gaze annotations for the person on the right

Explanation of annotation values:
'm': eye contact
'b': blink
'r': gaze to the right
'l': gaze to the left
'o': gaze upwards
'u': gaze downwards
'or': gaze to upper right
'ol': gaze to upper left
'ul': gaze to lower left
'ur': gaze to lower right
'na': annotation was not possible (i.e. because of strong motion blur)
'NaN': no annotation (occurs because the frames annotated for left and right person differ)

Download (140k)

In case of questions, please do not hesitate to contact Philipp Müller ().

License: CC BY 4.0

If you use this dataset in a scientific publication, please cite the following paper:

  1. Anticipating Averted Gaze in Dyadic Interactions

    Anticipating Averted Gaze in Dyadic Interactions

    Philipp Müller, Ekta Sood, Andreas Bulling

    Proc. ACM International Symposium on Eye Tracking Research and Applications (ETRA), pp. 1-10, 2020.

    Abstract Links BibTeX Project