Design and Evaluation of a Mobile Application for Real-Time Monitoring of Physical Exercises in Home-Based Cardiac Rehabilitation

This paper presents the design and preliminary evaluation of a mobile application developed to support home-based rehabilitation for patients recovering from cardiac surgery. The application uses the smartphone’s built-in RGB camera in conjunction with MediaPipe’s pose landmark detection to track body articulations and monitor exercise execution. A rule-based approach, driven by landmark positions, was implemented to evaluate user poses during the exercises. Based on this evaluation, the system tracks progression through a predefined sequence of states that represent correct execution. This method enables automatic repetition counting and the detection of common execution errors. Real-time visual and auditory feedback is provided, effectively simulating a virtual coach that guides patients through their rehabilitation routines. The application was tested with a specific rehabilitation exercise and demonstrated good reliability in both repetition counting and error detection. Although some limitations related to depth estimation by the smartphone camera were observed, the system shows promise as an accessible tool for guided rehabilitation, without the need for specialized equipment or clinical supervision, and has the potential to enhance patient engagement and adherence to rehabilitation protocols.