Patient Monitoring System using Computer Vision for Emotional Recognition and Vital Signs Detection

Patient monitoring is a pivotal part of the healthcare system nowadays, either at hospitals or at home. Critical patients require to be monitored consistently and less human involvement, 24 hours a day to enable them to get medical assistance in the moment of need. However, these types of services are only available in private hospitals. Typically, there is a small number of patients in private hospitals, especially from the higher socio-economic backgrounds. Conversely, in public hospital, a huge number of patients require medical attention due to the imbalance between staff to patient ratio. The patient monitoring system is restricted when they are asleep or unconscious due to incapability to call for assistance during an emergency. This may delay the treatment as the medical staff are unaware of the patients’ condition, hence resulting fatality. This work proposes a smart integrated patient monitoring system that automatically detects patient’s emotional state and heartbeat levels through face recognition algorithms, heartbeat and temperature sensors. A Raspberry Pi and NodeMCU are used as client nodes to collect the patient data. These data are then transmitted to an IoT cloud for realtime visualization. Through this monitoring system, critical patients can get immediate attention without the requirement of the staff being present there 24 hours a day. This system offers a faster response from medical staff to provide treatment in critical times.