Objective. Exercise monitoring provides valuable insights into the cardio-respiratory health and fitness performance of exercisers. To address the limitations of existing studies that only monitors specific phases of the fitness cycle, this study introduces a novel approach for camera-based monitoring throughout the entire fitness cycle, encompassing the pre-exercise, during-exercise, and post-exercise phases. Approach. Validated video-based algorithms were employed to monitor physiological parameters, including heart rate (HR), HR variability (HRV) (time-domain, frequency-domain and nonlinear-domain metrics), and respiratory rate (RR). Measurements were conducted using a camera positioned in front of a treadmill, along with electrocardiogram (ECG), PPG recorded simultaneously for benchmarking. This work comprised of a total of 36 adult subjects (18 males, 18 females; average age: 21.3 ± 2.8 years), which are categorized into subjects with regular exercise habits (ES) and those without (NS) (ES: 10, NS: 26) based on their performance of this running trial organized in our study. Main results. The results showed that the camera-based system performed well in HR, RR and HRV measurement. In the pre-exercise phase, camera-based monitoring achieved an mean absolute error of 2.74 bpm for RR and 12.19 bpm for HR. HRV parameters, including mean interbeat interval and very low frequency, showed Pearson correlation coefficients of 0.99 and 0.97, respectively, with ECG. Compared to NS, ES exhibited more robust cardio-respiratory functioning, characterized by lower HR during exercise and faster HR recovery during post-exercise. Camera-based monitoring effectively captured these differences in physiological parameters across the fitness cycle. Significance. This study validates the feasibility and effectiveness of camera-based monitoring throughout the full fitness cycle. The findings highlight the contrasting cardio-respiratory responses between ES and NS, emphasizing the potential of camera-based systems in providing comprehensive, non-invasive insights into exercisers' fitness performance and cardiovascular health.

The source code and dataset will be made open-source upon the acceptance at this site https://github.com/contactless-healthcare/Camera-based-Monitoring-for-Full-Fitness-Cycle.