An Estimation Model of Nasopharyngeal Specimen Collection Path for Non-face-to-face Automatic Sampling Robot

Abstract

Due to the worldwide spread of COVID-19, each government invests many human resources and money in screening tests. The spread of the virus has led to the development of robots that track the location of specimen collection or drive directly through master-slave devices by installing special equipment on patients' noses to reduce the physical burden on medical staff and prevent infection during screening tests. Sampling robots proposed in previous studies have a rather complicated specimen collection process or make it impossible to collect specimens when the patient cannot wear special equipment. Therefore, we propose a deep learning-based model that predicts the nasopharyngeal specimen sampling path without additional equipment. The test bench for the collection of learning datasets was configured, and the nasopharyngeal specimen sampling path was expressed using an augmented reality marker to learn the estimated value. In addition, we add weight factors to the proposed model to compare the root mean square error of the direction vector.