Please join this two part seminar. All are welcome to stay after the seminar for informal networking with the the center team.
MASTERS: Multisensory VR Environments to Train First Responders under Emotional Stress
The MASTERS project brings together a transdisciplinary consortium spanning computer science, HCI, mechanics, philosophy, cognitive sciences, and medicine to design innovative training methods for first responders (law enforcement officers, firefighters, emergency medical staff, etc.). Its central goal is to support them in maintaining self-control under high-stress conditions, whether the constraints are psychological, social, or situational.
Building on field studies, the project aims to develop a multisensory, immersive, and collaborative VR training program that realistically simulates socio-emotional situations. The presentation will first provide an overview of the project and its methodological approach before turning to the key challenges in designing such environments, including interpersonal haptic interactions, user representations, and interactions with virtual agents, elements that are critical to ensuring realism and training efficiency.
Speaker:
Amine Chellali received his M.Sc. in Robotics from École Centrale de Nantes (France) in 2005 and his Ph.D. in HCI from the University of Nantes (France) in 2009. From 2009 to 2011, he worked as a research assistant at the IRCCyN Lab. He then moved to the U.S. to work as a postdoctoral associate at Tufts University in the EREL Lab from 2011 to 2012 and then as a research associate in the ISSyL lab at Harvard Medical School. Currently, he is an Associate Professor in the Department of Computer Engineering at the University of Evry Paris-Saclay (France) and co-leader of the Interaction, Mixed Reality, and Assistive Robotics (IRA2) research group at the IBISC Lab. His research interests include Simulator fidelity, Haptic interactions, Collaborative virtual environments, and the transfer and evaluation of human skills. The main applications are in the field of medical training. He has received several grants for his research on VR simulation from the French National Research Agency and other national funding sources.
Emotion Analysis and Empathic Adaptation in Immersive Scenarios: Toward Ethically Informed Human Understanding and Engagement
This research first explores the use of adaptive immersive scenarios to enhance human understanding and engagement in different contexts. The models combine ontologies, Large Language Models (LLMs), and Fuzzy Cognitive Maps (FCM) to produce stressful scenarios for training first responders. These scenarios are personalized according to each user’s profile, accounting for individual vulnerabilities—such as empathy, sensitivity to authority, or stress triggers—to create ethically informed and effective experiences.
Building on this concept of user-centered scenario design, the ITS-STORY project investigates the use of mixed reality to foster historical empathy. Focusing on the Royallieu Camp Memorial in Compiègne, the project examines how immersive technologies can renew public engagement with the history of internment and deportation during the Second World War. Among the projects's objectives is the adaptation of immersive scenarios to participants' empathic responses, enabling personalized and ethically sensitive experiences that foster deeper historical understanding while avoiding exaggerated emotional engagement.
Within this framework, the research focuses on the detection and quantification of empathy through facial emotion recognition using computer vision and machine learning. The aim is to understand whether empathic responses—elicited by historically grounded virtual scenarios—can be inferred from emotional cues. Building on multiple iterations of facial emotion recognition models, the work has evolved toward explainable approaches, combining deep learning architectures with SHAP-based interpretability analyses.
This presentation addresses the adaptive design of scenarios according to user responses, details the methodological framework for measuring empathy through facial emotion analysis, and discusses ongoing challenges in linking emotional and empathic responses in immersive memorial environments. Together, these initiatives highlight the potential of personalized immersive experiences to enhance both professional training and public engagement and demonstrate how measuring emotional and empathic responses can inform ethically sensitive scenario design.
Speakers:
Domitile Lourdeaux is Full Professor at the CNRS Heudiasyc UMR 7253 laboratory at the Université de technologie de Compiègne UTC – Alliance Sorbonne University. Since 1997, she has been conducting research on Artificial Intelligence based Virtual Environments.
Since 2019, she has been coordinator of the Artificial Intelligence and Data Science Speciality at UTC.
She has coordinated and participated in numerous projects: H2020 INFINITY (European police cooperation for criminal investigations), 2020-2024 ANR VICTEAMS (2014-2019 training of teams to rescue injured people), MacCoy Critical (2014-2019 training non-technical skills in critical situations), and SAGECE (2008-2010 training in crisis management following NRBCE terrorist attacks) to name a few.
Insaf Setitra is an Associate Professor in the CNRS Heudiasyc UMR 7253 laboratory at the Université de technologie de Compiègne UTC – Alliance Sorbonne University. Before joining Heudiasyc, she worked as a Research Engineer at the Research Center on Scientific and Technical Information (2010 to 2019) and later became an Associate Professor at the University of Science and Technology Houari Boumediene (USTHB), Algeria (2019 to 2023). She was also a Visiting Associate Professor at Chubu University, Japan in 2023 and 2024, and previously served as an Invited Researcher at the National Institute of Informatics (NII) in Tokyo in 2013 and 2015. She holds a Ph.D. in Artificial Intelligence (2019), a master’s degree in software engineering (2010), a bachelor’s degree in information systems and software engineering (2008), all from USTHB, and a bachelor’s degree in finances from the University of Management and Economics of Algiers. Her research interests include computer vision, deep learning, and explainable AI and their applications to affective computing and autonomous vehicles.
Jeanne Parisse is a PhD student working on the MASTERS project. She holds an Engineering degree in Computer Science from Université de technologie de Compiègne UTC – Alliance Sorbonne University. Her research focuses on adaptive scenario modeling to generate personalized, stress-inducing scenarios.