Knowledge-Driven Dialogue and Visual Perception for Smart Orofacial Rehabilitation
Authors: Jacobo López Fernández Meritxell García Perea
Date: 12.12.2023
Abstract
This paper addresses the problem of accomplishing Orofacial Rehabilitation (OR) with the assistance of artificial intelligence. The main challenges involve accurately monitoring and interacting with the trainees, while preserving user experience. We analyse different approaches to solving these challenges and propose a methodology to build smart knowledge-driven OR systems that focus on automated interaction. Our proposal leverages the combination of vision-based micro and macro facial expression recognition and skill-based dialogue systems, which facilitate encapsulating the knowledge of rehabilitation professionals into natural language interactions. Experimental results of spoken keyword spotting and micro and macro facial expression recognition algorithms are provided. The OR expressions image dataset employed in our experiments is also published to support further research in the field.
BIB_text
title = {Knowledge-Driven Dialogue and Visual Perception for Smart Orofacial Rehabilitation},
pages = {397-411},
keywds = {
Dialogue Systems; Facial Expression Recognition; Orofacial Rehabilitation
}
abstract = {
This paper addresses the problem of accomplishing Orofacial Rehabilitation (OR) with the assistance of artificial intelligence. The main challenges involve accurately monitoring and interacting with the trainees, while preserving user experience. We analyse different approaches to solving these challenges and propose a methodology to build smart knowledge-driven OR systems that focus on automated interaction. Our proposal leverages the combination of vision-based micro and macro facial expression recognition and skill-based dialogue systems, which facilitate encapsulating the knowledge of rehabilitation professionals into natural language interactions. Experimental results of spoken keyword spotting and micro and macro facial expression recognition algorithms are provided. The OR expressions image dataset employed in our experiments is also published to support further research in the field.
}
isbn = {978-303134585-2},
date = {2023-12-12},
}
