Scripts to parse the provided data and to compare saliency maps and scanpaths are also available (in the FTP containing the Training Dataset).
In particular, the following metrics are provided:
- Comparing saliency maps:
- Linear Correlation Coefficient (LCC)
- Kullback-Leibler divergence (KLD)
- Normalized scanpath saliency (NSS)
- Area under the curve (AUC)
- Similarity Measure (SIM)
- Comparing scanpaths:
- Similarity metric based on MultiMatch (Jardozka et al.), with some modifications to deal with 360º content.
For more details and citation, please see:
- Jesús Gutiérrez, Erwan J. David, Yashas Rai, and Patrick Le Callet, “Toolbox and dataset for the development of saliency and scanpath models for omnidirectional/360° still images”, Signal Processing: Image Communication, 2018.
- Erwan J. David, Jesús Gutiérrez, Antoine Coutrot, Matthieu Perreira Da Silva, and Patrick Le Callet, “A Dataset of Head and Eye Movements for 360° Videos”, Proceedings of the 9th ACM on Multimedia Systems Conference (MMSys’18), Amsterdam, Netherlands, Jun. 2018.