SUPER - Solution for Uncertainty and personalization in Emotion Recognition

Abstract

MIT project MISTI Global Seed Fund : Collaboration MIT/LPNC(MIAI)/CEA.

Coordinators of the project

  • Martial Mermillod
  • Picard Rosalind
  • Matt Groh
  • Katherine Anne Matton
  • Marina Reyboz, Christelle Godin du CEA (LIST & LETI)
  • Marion Mainsant (CEA-LETI)

Description

This project addresses the topic of estimating mental states, including emotions, stress levels, and mental health illnesses, from data collected via sensors, cameras, and smartphones. Being able to measure mental states can help to improve people’s health, safety, and well-being. For example, during machine operation or semi-autonomous driving, detecting when an operator or driver’s capacities are impaired by distractions, drowsiness, or stress, would allow preventive action to be taken, like temporarily increasing the degree of machine automation or sending a warning signal to the user. In the case of e-learning or gaming, detecting mental load, boredom, and the emotions of the learner or gamer, would help the system/teacher to adapt the content favorably. In the case of mental health patients, detecting emotions, stress, or specific patterns of mental illness could accelerate diagnosis and enable timely interventions that help to avert crises.

Budget of the project

29 900 USD for 2 years.