SPEECH RECOGNITION SYSTEM BASED ON WORD STATE DURATION AND/OR WEIGHT

摘要

A low cost, speaker independent, limited vocabulary, word recognizing microcomputer (17). The microcomputer divides each spoken word into a series of word states, determines the length of each state and classifies each state as fricative, vowel-like, or silent. The incoming speech pattern, in the form of an array of classified word states and an array of associated word lengths is then compared sequentially with a series of templates, defining the limited vocabulary stored in the microcomputer's memory (19). Each stored template includes an array of classified states, an array of lengths corresponding to the states and an array of weights corresponding to at least some of the states. The state classifications of the incoming state array are compared sequentially with the template states of each template state array and where the classifications match, an error score is generated based on the difference in lengths between the template state lengths and the word state lengths. The weights are used to permit recognition of a spoken word as a template word even when the array of states representing the spoken word is not identical to an array of states in any of the template words.