Training When a speaker attempts to verify himself with this system, his incoming signal is compared to that of a "key". This key should be a signal that produces a high correlation for both magnitude and pitch data when the authorized user utters the password, but not in cases where: the user says the wrong word (the password is forgotten) an intruder says either the password or a wrong word To develop such a key, the system is trained for recognition of the speaker. In this instance, the speaker first chooses a password, and it is acquired five separate times. The pitch and magnitude information are recorded for each. The signal that matches the other four signals best in both cases is chosen as the key. Furthermore, the system will return lower thresholds for matching to magnitude vs. time and pitch vs. time. Thresholds below 0.90 are not returned, and to eliminate the possibility of an extremely high threshold like 0.99, an upper bound is placed on the thresholds of 0.95. Therefore, the thresholds will lie between 0.90 and 0.95. In MATLAB, the function makelock.m was written to determine the key signal from five possible signals. The call is: [locksignal,pitch_thresh,mag_thresh] = makelock(sig1,sig2,sig3,sig4,sig5) Example: In this instance, JP records his chosen password "project" five times and saves them as jp1, jp2, etc. When the call is made, the results are assigned to an array lock (which holds the time signal, a large array of points), and two scalars with the lower threshold bounds. >> [lock,pitchthreshold,magthreshold]=makelock(jp1,jp2,jp3,jp4,jp5); >> pitchthreshold pitchthreshold = 0.9500 >> magthreshold magthreshold = 0.9500 >> sum(jp1-lock) ans = 0 Hence, jp1 has been chosen as the key. << 8 || 10 >> APATHY - "If we don't take care of the customer, maybe they'll stop bugging us." © 1999Sara MacAlpineJP SlavinskyNipul BharaniAamir Virani