Method for building language model, speech recognition method and electronic apparatus

• US 9,711,138 B2
• Filed: 09/29/2014
• Issued: 07/18/2017
• Est. Priority Date: 10/18/2013
• Status: Active Grant

• Current Assignee: VIA Technologies Incorporated (VIA Technologies)
• Original Assignee: VIA Technologies Incorporated (VIA Technologies)
• Inventors: Zhang, Guo-Feng
• Primary Examiner(s): MCFADDEN, SUSAN IRIS
• Application Number:
  US14/499,261
• Publication Number:
  US 20150112679A1
• Time in Patent Office: 1,023 Days
• Field of Search: 704 9, 704 10, 704243
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/14 using statistical models, e.g. Hidden Markov Models [HMMs] G10L15/18 takes precedence
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/0633 using lexical or orthographic knowledge sources

Abstract

A method for building a language model, a speech recognition method and an electronic apparatus are provided. The speech recognition method includes the following steps. Phonetic transcriptions of a speech signal are obtained from an acoustic model. Phonetic spellings matching the phonetic transcriptions are obtained according to the phonetic transcriptions and a syllable acoustic lexicon. According to the phonetic spellings, a plurality of text sequences and a plurality of text sequence probabilities are obtained from a language model. Each phonetic spelling is matched to a candidate sentence table; a word probability of each phonetic spelling matching a word in a sentence of the sentence table are obtained; and the word probabilities of the phonetic spellings are calculated so as to obtain the text sequence probabilities. The text sequence corresponding to a largest one of the sequence probabilities is selected as a recognition result of the speech signal.

Claims

1. A method for building a language model, adapted to an electronic apparatus, the method comprising:

obtaining a text corpus through training with a plurality of speech signals based on different languages, dialects or different pronunciation habits;

receiving a plurality of candidate sentences; and

obtaining a plurality of phonetic spellings matching each of words in each of the candidate sentences and a plurality of word probabilities by training with the text corpus, so as to obtain a candidate sentence table corresponding to the candidate sentences,

wherein the step of obtaining the text corpus through training with the speech signals based on different languages, dialects or different pronunciation habits comprises;

receiving the phonetic spellings matching pronunciations of each of the words according to the corresponding words in the speech signals; and

obtaining the word probabilities of each of the words corresponding to each of the phonetic spellings in the text corpus by training according to each of the words and the phonetic spellings.

2. An electronic apparatus, comprising:

an input unit, receiving a plurality of speech signals;

a storage unit, storing a plurality of program code segments; and

a processing unit, coupled to the storage unit, the processing unit executing a plurality of commands through the program code segments, and the commands comprising;

obtaining a text corpus through training with the plurality of speech signals based on the speech signals of different languages, dialects or different pronunciation habits;

receiving a plurality of candidate sentences; and

obtaining a plurality of phonetic spellings matching each of words in each of the candidate sentences and a plurality of word probabilities by training with the text corpus, so as to obtain a candidate sentence table corresponding to the candidate sentences,

wherein the command of obtaining the text corpus through training with the speech signals based on different languages, dialects or different pronunciation habits comprises;

receiving the phonetic spellings matching pronunciations of each of the words according to the corresponding words in the speech signals; and

obtaining the word probabilities of each of the words corresponding to each of the phonetic spellings in the text corpus by training according to each of the words and the phonetic spellings.