Speech processing device and method, and program for recognition of out-of-vocabulary words in continuous speech

• US 7,653,541 B2
• Filed: 11/12/2003
• Issued: 01/26/2010
• Est. Priority Date: 11/21/2002
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Ogawa, Hiroaki
• Primary Examiner(s): Hudspeth; David R
• Assistant Examiner(s): Shah; Paras
• Application Number:
  US10/502,169
• Publication Number:
  US 20050143998A1
• Time in Patent Office: 2,267 Days
• Field of Search: 704/215, 704251-257
• US Class Current: 704/251
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
• CPC Class Codes: G10L 2015/022 Demisyllables, biphones or triphones being the recognition units
• CPC Class Codes: G10L 2015/027 Syllables being the recognition units
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering

Abstract

A speech processing device and a speech processing method, a storage medium, and a program decreases deletion errors and increases a speech recognition rate. A network of words and syllables is generated, and the network has two kinds of paths: paths that do not contain a particular syllable and paths that contain the syllable at a position corresponding to a boundary between words. Thus, an optimal sub-word sequence on the network is selected for an input utterance.

Claims

1. A speech processing device, including a processor, for processing an input utterance and registering a word contained in the input utterance based on the processing result, comprising:

recognition means for recognizing a continuous input utterance;

unknown word determination means for determining whether a result of recognition by the recognition means contains an unknown word;

network generating means for generating a network having two kinds of paths when the unknown word determination means determines that the result of recognition by the recognition means contains the unknown word;

wherein a first kind of path includes particular sub-words corresponding to word boundaries on one side or both sides of the unknown word, and

wherein a second kind of path does not include the particular sub-words,

wherein the particular sub-words are sub-words at positions where boundaries of the unknown words correspond to halfway points of the sub-words;

acquiring means for acquiring a word corresponding to the unknown word when the unknown word determination means determines that the result of recognition by the recognition means contains the unknown word; and

registering means for registering the word acquired by the acquiring means while associating the word with other information.

2. The speech processing device according to claim 1, further comprising pattern determination means for determining whether the result of the recognition matches a predetermined word sequence,

wherein the registering means registers the word if the result of the recognition matches the predetermined word sequence.

3. The speech processing device according to claim 2, further comprising reply generating means for generating a reply signal corresponding to the input utterance when the unknown word determination means determines that the result of the recognition does not contain an unknown word or the pattern determination means determines that the result of the recognition does not match the predetermined word sequence.

4. The speech processing device according to claim 2, wherein the registering means registers the word while associating the word with a category serving as the other information.

5. The speech processing device according to claim 2, wherein the registering means registers the other information while associating the other information with the matched pattern if the pattern determination means determines that the result of the recognition matches the word sequence.

6. The speech processing device according to claim 1, wherein the acquiring means acquires the word by clustering the unknown words.

7. The speech processing device according to claim 1, wherein the network generated by the network generating means is a network of words and syllables.

8. The speech processing device according to claim 7, further comprising:

selecting means for selecting a path having the highest score in the network based on matching between the input utterance and the network; and

pronunciation acquiring means for acquiring a pronunciation for the unknown word based on the network containing the path selected by the selecting means.

9. The speech processing device according to claim 1, further comprising:

comparison means for comparing an acoustic score obtained by matching a predetermined duration of the input utterance against known words with an acoustic score obtained by recognition of a syllable typewriter,

wherein the comparison means estimates that the duration corresponds to an unknown word if the acoustic score by recognition of a syllable typewriter is superior to the other score.

10. The speech processing device according to claim 9,

wherein the comparison means compares the acoustic score obtained by matching a predetermined duration of the input utterance against known words with the acoustic score obtained by recognition of a syllable typewriter after the comparison means corrects the acoustic score obtained by recognition of a syllable typewriter.

11. A speech processing method, utilizing a processor in a speech processing device that processes an input utterance and registers a word contained in the input utterance based on the processing result, comprising the steps of:

recognizing a continuous input utterance, utilizing a recognition unit;

determining whether a result of recognition in the recognition step contains an unknown word, utilizing an unknown word determination unit;

generating a network having two kinds of paths, utilizing a network generating unit,

wherein a first kind of path includes particular sub-words corresponding to word boundaries on one side or both sides of the unknown word, and

wherein a second kind of path does not include the particular sub-words when the determining step determines that the result of recognition contains the unknown word;

wherein the particular sub-words are sub-words at positions where boundaries of the unknown words corresponds to halfway points of the sub-words;

acquiring a word corresponding to the unknown word when the determining step determines that the result of recognition contains the unknown word, utilizing an acquiring unit; and

registering the word acquired in the acquiring step while associating the word with other information, utilizing a registering unit.

12. A storage medium for storing a computer-readable program executed in a speech processing device for processing input utterances and registering words contained in the input utterances, wherein the computer-readable program comprises the steps of:

recognizing a continuous input utterance;

determining whether a result of recognition in the recognition step contains an unknown word;

generating a network having two kinds of paths, when the determining step determines that the result of recognition contains the unknown word,

wherein a first one kind of path includes particular sub-words corresponding to word boundaries on one side or both sides of the unknown word, and

wherein a second kind of path does not include the particular sub-words,

wherein the particular sub-words are sub-words at positions where boundaries of the unknown words correspond to halfway points of the sub-words;

acquiring a word corresponding to the unknown word when the determining step determines that the result of recognition contains the unknown word; and

registering the word acquired in the acquiring step while associating the word with other information.