JPS58159595A - Monosyllabic voice recognition system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] (a) Technical field of the invention This invention relates to a monosyllabic speech recognition method for recognizing speech.

Background of the Technology As speech recognition technology has improved in recent years, there has been a desire for a speech recognition device with fewer recognition errors when recognizing a speaker's voice. The speech recognition method mainly converts the speaker's monosyllabic speech into feature parameters and stores them in advance, and then compares the feature parameters of the unknown input monosyllabic speech with the pre-stored feature parameters to find the one that is most similar. The system recognizes the corresponding monosyllabic speech, but even if the same monosyllabic speech is pronounced, the characteristic parameters change depending on the way it is uttered. It is difficult to reduce it to zero. In particular, for monosyllabic speech that has characteristic parameters that are likely to cause recognition errors, it is impossible to improve the recognition rate unless the matching method is taken into consideration. For this reason, after comparing all monosyllabic voices registered in advance with the monosyllabic voice of the speaker, multiple replays are performed in order from the monosyllabic voice that is most similar to the unknown input monosyllabic voice based on the matching results. The recognition rate is calculated by selecting matching candidates from the registered monosyllabic voices and re-matching the unknown input monosyllabic speech with the re-matching candidates for the Zen number using the re-matching parameters determined according to the combination of the plurality of re-matching candidates. A method has been proposed to measure the direction of the earthquake. However, there is room for improvement in the above-mentioned re-verification method, and countermeasures are desired. This aims to shorten the time required for town verification by narrowing down the number of candidates for reverification, and to improve economic efficiency by simplifying the configuration of the device.

(d) Structure of the Invention The structure of the present invention is best achieved by registering monosyllabic speech in advance, and querying the feature parameters of the unknown input monosyllabic speech and the feature parameters of all previously registered monosyllabic speech in the DP. A plurality of rematching candidates are sequentially selected from the registered monosyllabic speech in descending order of similarity, and the unknown input monosyllabic speech and the rematching candidate are selected based on a rematching parameter determined according to the combination of the plurality of ladle matching candidates. As a result, the most similar re-matching candidate is recognized as the corresponding monosyllabic speech, but when selecting the plurality of re-matching candidates, it is necessary to distinguish between the highest re-matching candidate in DP matching and the unknown. If the degree of similarity with the input monosyllabic speech is equal to or greater than the threshold that can be satisfied by the re-matching candidate of the -th rank, the re-matching step is omitted and the re-matching candidate of the -th rank is sent out as the recognition result, and the monosyllabic speech is output. This aims to shorten recognition time and simplify the re-verification circuit.

(e) Examples of the invention The figure is a block diagram of a circuit showing one embodiment of the present invention.

First, the speaker connects the switching section 3 to the parameter storage section 4 under the control of the control section 8 in order to register monosyllabic speech in advance, and inputs the monosyllabic speech.

The front rudder processing section 1 performs audio level adjustment, analog-to-digital conversion, etc., and sends it to the parameter extraction section 2. The parameter extraction section 2 extracts characteristic parameters of the monosyllabic speech and stores them in the parameter storage section 4. Next, in order to recognize a monosyllabic voice, the speaker selects the switching unit 3 under the control of the control unit 8.
is connected to the storage unit 5, and a monosyllabic voice is uttered. The characteristic parameters of the unknown input monosyllabic speech that have entered the storage unit 5 through the preprocessing unit 1, parameter extraction unit 2, and switching unit 3 through the same operation as described above are stored in the parameter 3-storage unit 4 under the control of the control unit 8. The matching unit 6 performs DP comparison with the feature parameters of all the monosyllabic speeches, and the monosyllabic speech with the most similar feature parameters among the feature parameters of all the monosyllabic speeches is designated as the highest re-matching candidate. Then, a plurality of re-verification candidates are sequentially selected and sent to the determination unit 7.

The determining unit 7 determines the degree of similarity based on the distance between the unknown input monosyllabic speech calculated by the matching unit 6 and the re-matching candidate. That is, the degree of similarity between the re-matching candidate selected as the first-ranked re-matching candidate and the unknown human monosyllabic speech by the matching unit 6 as described above is greater than the threshold predetermined by the second-ranking re-matching candidate, and the re-matching is performed. If it is determined that the candidate is almost definitely an unknown input monosyllabic speech, it is sent as an output through the control unit 8 as a recognition result. However, if the degree of similarity between the re-matching candidate at the highest rank and the unknown input monosyllabic speech is smaller than the threshold value and there is a danger in determining the re-matching candidate at the highest rank as the unknown input monosyllabic speech, re-matching is necessary. Recognize actions step by step. Therefore, the control unit 8 transfers the feature parameters corresponding to the re-verification candidate from the parameter storage unit 4 to the multiplier 10, and transfers the feature parameters of the unknown input monosyllabic speech stored in the storage unit 5 to the multiplier 11. The determining unit 7 determines the rematching parameters determined by the rematching candidates, that is, emphasizes the frequency band components suitable for mutually discriminating the rematching candidates, and reduces other frequency band components as frequency weights. Storage part 1
2 to the multipliers 10 and 11. The 0-or determination unit 7 sends the threshold value, which is a parameter for determining the optimal matching interval determined according to the re-matching candidate, from the threshold storage unit 13 to the re-matching unit 9. The combination unit 9 performs re-verification using the output of the multiplier 10.11 and the threshold value from the threshold value storage unit 13. A plurality of re-verification candidates are re-verified with the unknown input monosyllabic speech in order from the -th rank re-verification candidate. The most similar re-verification candidate after collation is sent to the output from the control unit 8 as a recognition result (f) As described in detail, the present invention provides a monosyllabic speech recognition method using the re-referral method. The effect is significant because it is possible to reduce the number of reverification candidates, shorten the time required for reverification, and simplify the component equipment related to reverification operation, improving economic efficiency. There is.

[Brief explanation of drawings]

The figure is a block diagram of a circuit illustrating one embodiment of the present invention. 1 is a preprocessing unit, 2 is a parameter extraction unit, 3 is a switching unit,
4 is a parameter storage unit, 5 is a storage unit, 6 is a collation unit, 7 is a determination unit, 8 is a control unit, 9 is a re-verification unit, 10.11 is a multiplier, 12 is a frequency weight storage unit, 13 is a threshold value storage Department.

Claims (1)

[Claims] After comparing all previously registered monosyllabic speech with the unknown input monosyllabic speech, a plurality of re-matching candidates are selected from the registered monosyllabic speech based on the matching results, and a combination of the re-matching candidates is selected. In a speech recognition device that selects a rematching parameter for each case and rematches it with an unknown input monosyllabic speech, the similarity of the rematching candidate that ranks highest when selecting the rematching candidate is the rematching parameter. A monosyllabic speech recognition method that is characterized by omitting re-verification if the threshold value determined for each candidate is exceeded.