JPS6338720B2 - - Google Patents

Description

[Detailed description of the invention]

(Industrial Application Field) The present invention calculates the degree of similarity between the recognized phoneme sequence of input speech and the dictionary phoneme sequence of each dictionary entry in a word dictionary in which phonemes are expressed, using a confusion matrix. The present invention relates to a word speech recognition method for recognizing words. (Structure of conventional example and its problems) A conventional word speech recognition method will be explained with reference to FIG. As shown in FIG. 1, input speech is first analyzed, features of the input speech are extracted, and phonemes making up the input speech are recognized. This recognized phoneme sequence is compared with the dictionary phoneme sequence of each dictionary entry in the word dictionary, and the likelihood between the two phoneme sequences is calculated as follows:
It is calculated by finding the recognition probability for each phoneme using a Confusion Matrix (hereinafter abbreviated as CM) between phonemes, and the recognition word is determined by finding the dictionary entry that has the maximum likelihood between phoneme sequences. That is. Table 1 shows an example of a word dictionary used in the word speech recognition method, and each word is written according to the phoneme notation shown in Table 2. FIG. 2 shows a part of the CM. In Figure 2, vertical is a single

【table】

【table】

Claims (1)

[Claims] 1. The likelihood between the recognized phoneme sequence of the input speech and the dictionary phoneme sequence of each dictionary item in the word dictionary with phoneme notation indicates the probability of what kind of phoneme each phoneme in the word dictionary will be recognized as. When recognizing words by calculation using a conflation matrix, the conflation matrix is created using the analysis results of word audio data or the phoneme recognition results for which the correct answer is known in advance, and the confusion matrix is For elements with a low probability of occurrence, if the element cannot occur due to the nature of the voice, a value that can be considered as zero or zero is given to the probability, and if the element can occur due to the nature of the voice, a value that can be considered zero is given. A word speech recognition method characterized by using a conflation matrix that is given a finite value that is close to zero but not zero.