This report deals with noninvasive imaging of airway geometry based upon information contained in acoustic reflections measured at the mouth. Here we describe a new theoretical approach that enables development of a new miniaturized apparatus. Unlike the single-transducer systems used currently, this new strategy is based upon a two-transducer system that is a variant of that suggested originally by Shroeder (1967). We have developed, implemented, and tested computational algorithms necessary to reconstruct airway dimensions from acoustic reflection data using this two-transducer strategy.