Speech recognition system and method for recognising given speech patterns, specially for voice control

摘要

An HMM speech recognition device (1) stores preset speech samples such as possible commands for a speech-controlled car telephone. These samples are compared with a recorded speech expression. A probability calculation block (2) detects probabilities (P) and matches the recorded speech expression to any preset speech sample. The sample with the greatest probability (Pmax) for a match is selected and passed to a decision block (3).