Thèse - Using machine learning to capture & model the singularity of gaze and spatial trajectories
Thématique : machine learning,big data,eye-tracking, spatial navigation,video game,individual difference
Thématique : Signaux, Images, Ergonomie et Langues
Equipes : IPI - Image Perception Interaction
Etablissement : Ecole Polytechnique de l'Université de Nantes ( Polytech Nantes )
Description du poste :
Context :Understanding how people navigate and explore the world is critical to many scientific fields, including image processing, clinical diagnosis and cognitive states inference. In the last few years, many studies proposed models of visual attention trying to predict what in a scene attracts the observers’ attention. However, models’ prediction are still far from human behavior.
Project :In this PhD project, the candidate will use machine learning to capture observer’s singularity in how they explore the world. By capturing and modelling these singularities, the candidate will build models tailored for specific subgroups of the population (e.g. patients). The candidate will work with a planetary-wide database of over 4 million spatial trajectories crowd-sourced via a video-game, Sea Hero Quest. This constitutes one of the largest database of human behavior designed for scientific purposes in the history of science. The portability of the developed approches from gaze to spatial navigation trajectories will be studied.
- around 2025 €/month with teaching responsibilities).
- Funds are secured for travels to conferences in exotic locations.
- Starting date is flexible, between 01/01/2019 and 01/09/2019