Abstract This paper presents a novel approach for remotely detecting and monitoring a user's breathing rate in real-time using a high precision, single-point infrared sensor. Remote breathing detection is especially useful for rehabilitative robotic applications such as post-stroke and post-operative cardiac therapies since a patient's physical stress level can be continuously monitored and used to determine the appropriate level and duration of physical exertion throughout the course of therapy. Additionally, socially assistive robots which discretely collect breathing information from their users, can record physiological changes in real-time and customize interactions during the session based on the perceived stress state of the patient. The proposed approach is an important potential improvement for therapies where user mobility is an inherent part of the therapy or when users have a general aversion to being fitted with sensors. Further, due to its relatively small size and modular design, this system can be integrated into existing rehabilitative robot systems in order to enhance their functionality. This research delivers a new technique for capturing changes in the sub-nasal skin surface temperature to monitor breathing events. Temperatures are obtained by tracking the sub-nasal region of the face and continuously targeting and sampling the infrared sensor. The breathing rate is automatically extracted using a sinusoidal curve-fitting function which provides an estima ted rate in breaths per minute. Results from preliminary tests show this system effectively captures breathing rates within an error rate of under 2 breaths per minute in excess of 70\% of typical test cases.