AI-Based Mobile Application for Personalized Monitoring of Healthy Sitting Posture

Coşkun H.

in: AI-Driven Personalized Healthcare Solutions , Houssem Chemingui,Meriam Lamloumi, Editor, IGI Global yayınevi, Pennsylvania, pp.1-25, 2025

  • Publication Type: Book Chapter / Chapter Research Book
  • Publication Date: 2025
  • Publisher: IGI Global yayınevi
  • City: Pennsylvania
  • Page Numbers: pp.1-25
  • Editors: Houssem Chemingui,Meriam Lamloumi, Editor
  • Kütahya Health Sciences University Affiliated: Yes

Abstract

In modern workplaces, many individuals spend extended periods sitting, often in positions not recommended by health professionals, which can lead to skeletal, back, muscle, and heart issues. This study developed a mobile application that monitors sitting postures based on health guidelines. Using a smart seat cover embedded with electro-textile sensors from previous research, sitting experiments were conducted to assess posture. Participants were first instructed in healthy posture, then asked to sit as they preferred, including the healthy posture, for 5 minutes. The recorded data were categorized manually by reviewing experiment videos, creating a dataset of healthy and unhealthy postures. This dataset was classified using ANN, Gradient Boosting (GB), AdaBoost, and Random Forest algorithms, with a 70/30 train-test split and 5-fold cross-validation. ANN, GB, AdaBoost, and Random Forest achieved accuracy rates of 99.8%, 100%, 99.2%, and 99.3%, respectively. Based on the GB model, a Java-based Android mobile app was developed for real-time monitoring and posture notifications.