Non-contact tidal volume measurement through thin medium thermal imaging

Abstract

Abstract Tidal volume (TV) is frequently measured by pulmonologists to assess the breathing ability and respiratory behaviors of a patient. The most common method of measuring tidal volume, spirometry, generally requires the patient to breathe through a tube while wearing a nose clip. This method of measurement is uncomfortable for the patient, and alters their natural breathing behaviors. In this paper, we describe and implement a new approach to respiration monitoring and tidal volume estimation that preserves patient comfort while maintaining measurement accuracy. In this method, a patient breathes onto a thin medium placed perpendicular to the patient׳s exhaled airflow while a thermal camera records the heat signature on the opposite side of the medium. Measurements such as tidal volume, respiration rate, and nose to mouth distribution are extracted from these thermal images, correlated with ground-truth measurements and trained into a recurrent network model using TensorFlow. This method is comfortable for the patient, works for a variety of different body types and ages, and can be used in the comfort of an out-patient office.