1. The computerized analysis of spoken words in order to automatically respond to voiced commands: speech is analyzed by finding patterns in the spectrum of incoming sound and comparing them with stored patterns of elements of sound, as phonemes, or of complete words. Neural network techniques may also be used to convert spoken words into computer text.
2. This can also mean to analyze a person's voice in order to identify the speaker as really who s/he claims to be, as in security systems.
Also called Voice Recognition, esp. as it pertains to speaker validation for security.