A Contactless Healthcare System with Face Recognition

Nowadays, healthcare is becoming a new issue increasingly that worthy to care. Healthcare measurement methods can be categorized into two main areas: contact and contactless. Among them, contactless has more flexibility and comfort. In this paper, we combined the pulse rate measurement and face recognition based on deep learning into an embedded system. The system can help people measure their pulse rate in real-time after recognize the face. Our system contains a graphics processing unit (GPU) and central processing unit (CPU) co-design platform in order to processing face recognition in real-time. The experimental results show the recognition average accuracy is 98%. In pulse rate detection aspect, the provided mean absolute error (MAE) and root-mean-square error (RMSE) are 5.095 BPM and 5.976 BPM. The results reveals that our system can truly apply in daily life. In the future, we hope our system can be utilized in several scenarios such as company door-access, driving monitor, industry, etc.