Hybrid Posture Monitoring System Using Computer Vision and Sensor-Based Analysis

Prolonged improper sitting posture is one of the leading ergonomic issues associated with musculoskeletal disorders and improper spinal alignment. Early identification and constant correction are critical towards ensuring proper spinal alignment and alleviating fatigue. For this purpose, this paper proposes a hybrid posture analysis system that utilizes both force-based sensing and computer vision. The proposed work will make use of force-sensitive resistor sensors connected to an ESP32 microcontroller for identifying pressure variations in seating positions, in addition to a vision module utilizing a webcam and OpenCV with MediaPipe Pose for identifying improper upper limb sitting posture positions such as slouching and neck bending. The solutions derived from these approaches will be processed using a Python programming language setup. The future scope of this work will incorporate an intelligent posture predictor with cloud analysis support.