This paper aims to open a new line of research by demonstrating the feasibility of monitoring breathing function in a highly automated and noncontact fashion. The method is based on the statistical modeling of dynamic thermal data captured through a highly sensitive infrared imaging system. This method achieved an accuracy of 96.43% on a small set of subjects during rest and after brief exercise. It has the potential to provide a unique capability for sustained monitoring of chronic or acute breathing problems and in sleep studies by overcoming the deficiencies of the existing measurement modalities. The method also opens the way for the next generation contact-free polygraphy that will not affect the subject's psychophysiology.