| 摘要 |
PURPOSE: A method and an apparatus for character input by hybrid-type speech recognition is provided to exactly recognize user's speech by receiving a second speech of phonemic unit corresponding to a speech recognition error and by learning speech recognition of the second speech. CONSTITUTION: If an input mode is a soft consonant input mode, a control unit(102) makes a first voice processor(101a) recognize a received voice speech signal as a continuous speech of the first speech. If the input mode is a phoneme input mode, the control unit makes a second voice processor(101b) recognize the speech signal as a phoneme unit voice form of the second speech. The control unit inputs characters of a user by combination of speech recognition results. [Reference numerals] (100) Voice receiving unit; (101a) First voice processor; (101b) Second voice processor; (102) Control unit; (104) Display unit; (105) Storage unit; (107) Remote controller receiving unit
|
