Performance Assessment of Dynaspeak Speech Recognition System on Inflight Databases

摘要

Automatic speech recognition has demonstrated the potential to provide a natural interface to computing systems. For the military, the realization of this natural interface will depend on a robust speech recognition technology that is capable of handling the degraded speech conditions typical of the military aircraft environment. To aid in the assessment of various commercially available speech recognition systems, several aircraft speech databases have been developed at the Air Force Research Laboratory's Human Effectiveness Directorate. These databases attempt to capture the degraded audio conditions typical in the military airborne environment. This report discusses the results of an extensive evaluation of the DynaSpeak v1.4. 1 speech recognition system developed by Stanford Research Institute (SRI) using three different aircraft speech databases to benchmark its performance.