摘要 |
The present invention relates to a speech processing device and a speech processing method, a storage medium, and a program that can decrease deletion errors and increase a speech recognition rate. With reference to Fig. 16C, a network of words and syllables is generated, in which a path 91 that does not contain a syllable "WA" and paths 92 and 93 that contain the syllable "WA" at a position corresponding to a boundary between a word "WA" and a word "<OOV>", and a path 101 that does not contain a syllable "WA" and paths 102 and 103 that contain the syllable "WA" at a position corresponding to a boundary between a word "<OOV>" and a word "DESU". Consequently, an optimal sub-word sequence on the network can be selected for an input utterance. The present invention can be applied to a speech recognition system. <IMAGE>
|