The artificial utterance device which utilizes the biomedical signal

摘要

P To provide an artificial vocal apparatus easily controllable using myogenic potential, brain wave, motion of the limbs (acceleration), exhalation, contact pressure and or the like. PSOLUTION: This vocal apparatus comprises an input part for biological signals (EMG signals); a feature extraction processing part for extracting a feature pattern from the selected biological signal 1; a word sound estimating part for calculating probability ON a word sound intended by a user from the extracted feature pattern 2; a word estimating part for determining a word intended by the user from the time series change of posterior probability 3 ON the word sound calculated by the word sound estimating part; a sentence editor part for composing a sentence from the estimated word 4; a voice reproducing part for synthesizing voice and vocalizing based ON the prepared sentence information 5 and a feature amount ON sound volume, sound pitch and tone quality and vocalizing speed; a feedback part presenting the processing situation 7 of each part during the use of the apparatus to a user; and a parameter storage part capable of storing and reading parameters 8 used at each part. PCOPYRIGHT: (C)2005 and JPO& NCIPI