Designing Adaptive Immersive Therapeutic Spaces Using Convolutional Neural Networks for Community-Based Elderly Care

Abstract

As global aging accelerates, it is projected that by 2050, over 22% of the global population will be aged 60 or older. In this context, promoting healthy aging is crucial, with mental health challenges among the elderly receiving increasing attention. Art therapy has been recognized as an effective intervention to maintain physiological, cognitive, and social functions in older adults, significantly improving emotional well-being and social engagement. However, traditional art therapy faces limitations in scalability and assessment accuracy. Digital art therapy, with its real-time adaptability and sustainability, offers promising potential for expanding mental health interventions. This study focuses on Wuhan, China, surveyed and interviewed 2,000 elderly individuals to explore psychological health factors. Results indicate that individuals aged 60-69 exhibit significantly higher levels of depression and anxiety compared to those aged 50-59, with depression (M60-69=12.231>6.272,p<.001) and anxiety (M60-69=13.837>6.441,p<.001). High social participation was found to significantly enhance mental health (p<.001), with most respondents holding favorable views on community-based art therapy activities. To address these findings, this research proposes adaptive immersive therapeutic spaces for elderly care, integrating sensor and projection technologies. Using Convolutional Neural Networks (CNNs), the system analyzes real-time behavioral and emotional data, dynamically adjusting the visual and auditory elements to match the users' emotional states. A follow-up survey with 1,897 participants confirmed the feasibility of these spaces, with the majority anticipating improved emotional regulation. This study contributes to the advancement of mental health interventions for the elderly, offering novel perspectives for future research and practice.

Keywords: Digital Art Therapy, Healthy aging, immersive therapeutic space, Community-Based Elderly Care, Convolutional Neural Networks (CNNs)

DOI: 10.54941/ahfe1005885