ABSTRACT
Clinical data monitoring and storing are essential components of continuous and preventive healthcare systems. Data such as blood pressure, pulse rate, temperature, etc., can be recorded by the hospital staff daily for in-patient subjects. The usual way of noting them down is to check different parameters using various medical instruments and write it on paper with the corresponding patient's details (e.g., name, patient-id, or government identity card number). However, after the outbreak of COVID-19, there is a set of World Health Organization (WHO) guidelines to behave in public places. Ordinary people and professionals feel hesitant to touch any media even if they have some protection such as gloves and sanitizer. In this crisis, there is a natural demand for contact-less activities instead of touch-based traditional ways. Gesture-based activities might be one of the low-cost alternatives to some sensor-based systems. This paper uses a profound learning-based finger point gesture to capture writing in the air and realize it on the screen through a predictive model. Here, the proposed framework has been demonstrated as a proof of concept to record blood pressure data for multiple patients without touching any electronic screen or paper. The proposed architecture is developed based on the gesture recognition and metric learning, which have been used to recognize different digits trained from the MNIST digit dataset. The mean test accuracy is reached 99.47% on the same dataset. © 2022 IEEE.