Convolution Neural Network (CNN) Based Posture Detection System to Prevent Hunched Posture Syndromes

This paper presents a solution addressing the challenges associated with increased screen time during remote work. Leveraging specialized hardware and convolutional neural networks, our real-time postural detection system concentrates on monitoring the postures of the neck, shoulders, and arms. It offers recommendations to prevent health issues resulting from poor ergonomics. With an accuracy exceeding 80%, our low-power solution comprises a hardware device equipped with an LCD and buzzer for instant posture alerts, complemented by online notifications for prolonged bending. The ultimate goal is to advocate for a healthier remote working environment.