i-Care: A Multi-Modal Data Integration Approach for Real-time Surveillance and Voice Assistance to Improve Infant Care

Abstract

The rapid evolution and advancement of technology in adult healthcare have paved the way for improving infant care through enhanced monitoring and assistance systems. Traditional infant monitoring systems face challenges, including the need for large labelled datasets, high computational complexity, and limited real-world deployment due to resource constraints. This work presents i-Care, a novel multi-modal data integration approach designed to enhance real-time surveillance and voice assistance for improving infant care. i-Care addresses the traditional limitations by integrating vision, acoustic, and sensor data to provide a comprehensive and efficient monitoring solution. Our approach leverages deep learning techniques to seamlessly combine data from multiple sources, ensuring robust performance. i-Care not only improves the accuracy of infant activity recognition but also facilitates real-time voice assistance, aiding caregivers in promptly addressing the needs of infants. Experimental results demonstrate that i-Care outperforms existing methods in terms of accuracy, responsiveness, and adaptability, marking a significant step forward in infant care technology. © 2024 Copyright held by the owner/author(s).