- Project
- Completed
Game PLA[AI] – Real-time analysis of participation in games using artificial intelligence
François Johanny Engineer of Social Sciences
François Johany is a research engineer of social sciences studies at the National Institute of Agronomic Research (INRAE). He is interested in technological approaches for the evaluation of participatory processes, and board games in particular. He specializes in video analysis tools and manages the GAMAE platform’s digital technical platform.
-
Project start date :
2023/01/02 -
Status :
Completed -
Research organization :
INRAE -
Team :
Dr. Sylvain Dernat, Dr. Jocelyn de Goer, Dr. Séverine Bord, Simon Sayegh (INRAE)
In 2022, Game in Lab selected François Johany’s project to use artificial intelligence to provide the scientific community and board game publishers with a powerful method to analyze game session data. This project serves as a starting point for the development of complete capture and scientific analysis solutions for gaming sessions in a natural context.
Project overview
Although games are at the heart of new interfaces between science and society, their impacts on players can be difficult to analyze. Board games, in particular, generate multiple forms of data (verbal and non-verbal interactions between players, interactions with the game, understanding and application of the rules, etc.), with the volume and complexity of this data creating many analytical constraints. Currently, no technical tool is able to quickly identify the structuring elements of the course of a board game session. The objective of this research project is to assess the effectiveness of a complete data analysis solution (software and hardware) based on artificial intelligence (AI) to meet the scientific needs to analyze board game sessions.
![Game PLA[AI] – Real-time analysis of participation in games using artificial intelligence](https://cdn.svc.asmodee.net/production-gilv2/uploads/2024/08/Illustration-Francois-Johany-projet.webp)
Methodology
The protocol consists of several phases:
- An AI model capable of performing video observation analyses of game sessions is developed. Videos of existing game sessions are used to train the model to recognize recurring objects (hands, game pieces, etc.) and identify them according to particular elements (teams, organization, playing time, etc.).
- An ideal video capture environment is also tested.
- Game sessions are then held with children, teenagers or young adults to test the data produced by the AI model.
Outcomes
The results of this study are pending. This research represents a first step in overcoming the current limits of board game session analysis. A complete AI model could ultimately be included in discourse analysis, or even biometrics, and a mobile environment for capturing gaming sessions could be developed.
