MOVI-REACT

PR FESR 2021-2027 – OS 1.1 – Azione 1.1.1

ABSTRACT
MOVI-REACT is an R&D project aimed at developing an advanced platform for monitoring and assessing human movement in rehabilitation through the integration of computer vision, multimodal sensing, and Explainable Artificial Intelligence (XAI). By combining video analysis with data from wearable sensors and interpretable AI models, the project aims to create a patient-specific digital twin, enabling a deeper understanding of motor performance, rehabilitation progress, and future recovery trajectories. The platform is designed to support healthcare professionals with objective, data-driven insights while fostering trust through transparent AI-based recommendations. Potential applications extend beyond rehabilitation to sports performance analysis and industrial ergonomics.

Accurate assessment of movement quality is a critical component of rehabilitation pathways. Traditionally, therapists evaluate the execution of rehabilitation exercises through direct observation or video review. While effective, these methods can be time-consuming and subject to variability and subjectivity among evaluators.

Recent advances in computer vision, artificial intelligence, and wearable technologies have opened new opportunities for the automated and objective analysis of human movement. Techniques such as pose estimation, action recognition, and explainable AI make it possible to extract quantitative and clinically meaningful information, supporting decision-making and continuous monitoring of patient outcomes.

MOVI-REACT addresses these challenges by developing an integrated platform that combines data from wearable devices—including accelerometers, inertial measurement units (IMUs), and electromyography (EMG) sensors—with features extracted from video streams. This multimodal approach provides a comprehensive and robust characterization of movement patterns.

A key innovation of the project is the development of a patient-specific digital twin: a dynamic virtual representation capable of modelling motor behavior, tracking rehabilitation progress, and supporting predictive analyses. By integrating real-world observations with AI-driven models, MOVI-REACT aims to enable more personalized, objective, and efficient rehabilitation pathways, contributing to the next generation of intelligent healthcare solutions.

OUTCOMES
The MOVI-REACT project will deliver the following outcomes:

• A system for quantitatively assessing a patient’s rehabilitation pathway, thereby strengthening the evaluation of therapies and rehabilitation strategies.
• Computer vision algorithms designed to assess movement quality, including walking and standing up, with potential application across multiple contexts.
• A wearable sensor setup designed to support the analysis of human movement.
• A 3D digital twin that enables clinicians to better understand movement quality and manage the rehabilitation process more effectively.

PARTNERS
Rulex Srl (coordinator), BK Srl, SITEM Srl, National Research Council of Italy.

FUNDING BODY
The project was funded by Regione Liguria under the PR FESR 2021–2027 program – OS 1.1 – Azione 1.1.1.