Automatic Speaker Recognition over Military Communication Systems. A Feasibility Evaluation

摘要

Automatic speaker recognition (ASR) offers potential benefit for numerous Navy situations, including indentification of users of communication channels such as the telephone and channels using processed or vocoded speech. Currently the user must subjectively determine whether the person on the other end of the line is who he or she claims to be. However, past research has shown that ASR systems are capable of higher recognition accuracy than human listeners under certain circumstances. This report discusses a series of tests conducted to evaluate the feasibility of performing ASR using vocoded speech. The analog outputs of six different Department of Defense voice processors were used as input to a real-time ASR system. Data transmission rates of these processors ranged from 2400 to 64,000 bits per second. Recognition accuracy results for the processed speech were 70 to 95% using a 2500 Hz bandwidth input filter, and 75 to 95% using a 4000 Hz input filter. These results indicate that ASR using vocoded speech is definitely possible, though further research is needed to determine which speech parameters are best suited for use with each voice processor.