Survey on Sitting Posture Detection and Correction

Authors- Mr. Siddhant Bhalerao, Mr. Karan Birwadkar, Mr. Vedant Jadhav, Dr. Sunil Chavan

Abstract-Prolonged poor sitting posture can lead to health issues, highlighting the need for systems that encourage healthy habits. This project presents a human sitting posture detection and correction system using the Yolov5 model for real-time posture monitoring via a webcam. The system classifies user postures based on ergonomic criteria and provides corrective alerts when bad posture is detected. Preprocessing techniques, such as scaling and augmentation, are applied to the dataset to ensure model robustness. A user-defined timer is also integrated to remind users to take breaks. The trained Yolov5 model is evaluated for accuracy and implemented in a real- time detection system, with practical applications in promoting better sitting habits in home and office environments.

DOI: /10.61463/ijset.vol.12.issue6.321