HealCam: Energy-efficient and privacy-preserving human vital cycles monitoring on camera-enabled smart devices

Abstract In this paper, we present HealCam, an energy-efficient and privacy-preserving human vital signs (e.g., respiration cycles) monitoring system on camera-enabled smart devices. HealCam incorporates the related theories of compressive sensing in its system to reduce the sampling rate while preserving data privacy. HealCam saves significant cost on video processing via low-rate and non-uniform random sampling. It also provides a privacy-preserved data collection and enquiry service via lightweight compressive encryption and decryption scheme. According to our evaluations on real datasets, HealCam achieves high accuracy on respiration cycles reconstruction with extremely low average frame rate, i.e., 1 FPS, via non-uniform random sampling compared with traditional uniform sampling strategy. Then we implement HealCam on smartphones to evaluate its resource consumption. The results show that, HealCam is 23.6 times more energy efficient and 26.7 times faster than the original approach on video processing. Its data encryption component is 172 times faster while consumes only 1.01% energy compared with the corresponding state-of-the-art.