Wireless Device for Physical Activity Monitoring Tracking and Suggestion Using Image and Video Processing

As the need for advanced technologies in health care monitoring systems increases, this study work proposes the construction of a new wireless device that can identify physical activities in real time and provide personalized feedback. In contrast to other fitness trackers that use inertial sensors, our system utilizes image and video processing methods, thereby increasing accuracy further. The device comprises of low power camera module, microcontroller, and embedded machine learning algorithms that identify and classify human actions/ behaviors to be in one of the following classes: walking, running, squatting, sitting, or not active at all. The system applies computer vision tasks such as pose estimation, and uses learned activity profiles to recognition movement patterns associated with these activities. The device is aimed at providing users with personalized recommendation to improve their level of daily activities as well as their posture for better preventive health. Testing the system in different environmental settings with outdoor light and motion gave more than 92% accuracy in recognition and reliable wireless transmission of data. The device can be customized for different user profiles and activity levels, making it suitable for both fitness and rehab. Its modular design allows easy connection to the cloud for storing data and tracking progress. A mobile app offers feedback, shows trends, and gives personalized advice. This combination helps the device motivate people to be more active and take action to avoid inactivity.