Background: Maintaining the natural alignment of the spine during sleep is directly linked to musculoskeletal health, and the firmness of the mattress and pillow play a key role in this process. However, previous studies have largely relied on simple statistical indicators and have been limited in their precision of evaluating the balance of pressure distribution. Objectives: This study aimed to objectively evaluate the lumbar and cervical support of air-web material mattresses and pillows using AI technology and to identify optimal firmness parameters. Methods: Thirty healthy adults (15 men and 15 women aged 20~60 years) participated in the study. Body pressure distribution was measured using a high-resolution pressure mapping sensor (26×62, 1,612 sensels) for five mattresses (density 50~90 kg/m3) and three pillows. We defined selective support and two novel indices that incorporate biomechanical limits: the Lumbar Support Index (LSI) and the Cervical Support Index (CSI). The performance of five machine-learning models was evaluated for classifying high-support conditions. Results: Among the five tested mattresses, Mattress C (70 kg/m3) demonstrated the highest support, with a support score of 58.19, while Pillow 2 showed the best performance in terms of cervical alignment. Quadratic regression analysis estimated the optimal density at 74.74 kg/m3, indicating that the “medium–hard” category is optimal. Finally, the AI models achieved strong performance (ROC-AUC=0.84), suggesting that the LSI is intuitive and grounded in explainable physical principles. Conclusions: This study proposes new support indices based on body pressure distribution and identifies the 70~80 kg/m3 range as a “golden zone” that provides appropriate support for most users. These findings offer a new framework for evaluating sleep products using objective data and AI-based analysis and may serve as foundational evidence for developing personalized sleep solutions in the future.
Ⅰ. 서 론
Ⅱ. 재료 및 방법
Ⅲ. 결 과
Ⅳ. 고 찰
Ⅴ. 결 론
References
(0)
(0)