A vocabulary database construction method and the corresponding hunting and comparison method for voice identification system

Abstract

The invention relates to a method of building a character stock in a speech recognition system and a searching and comparing method thereof to solve the problem of calculating the same character repeatedly and reducing the whole operand. The method comprises the following steps of: 1) providing the data of polyphonic characters; 2) typing in the data; 3) building an acoustic model; 4) storing the data and the corresponding acoustic model thereof into the character stock. The character stock in the invention has the function of polyphonic characters recognition, so that the speech recognition system is closer to the pronunciation habit of the average users with human elements, therefore, enabling the users to follow the conventional pronunciation and to receive the correct recognition result.

Description

technical field [0001] The present invention relates to a vocabulary database construction method and its search comparison method for a speech recognition system, in particular to a vocabulary database construction method that can support polyphonic word processing and a more efficient search comparison method. Background technique [0002] The conventional speech recognition system does not include the processing function of polyphonic characters, so that users must pronounce another pronunciation of the polyphonic character to be able to recognize successfully when performing voice input. For example, the word "Xing" in the name Chen Lixing Must be pronounced as "plant '" to be successfully recognized, such as the user pronounced " '" can not be recognized correctly. For example, the word "乐" in the orchestra must be pronounced as " `" can be recognized, if it is pronounced as "ㄩ `" cannot be recognized correctly, and this voice input method is very different f...

AI Technical Summary

Problems solved by technology

[0002] The conventional speech recognition system does not include the processing function of polyphonic characters, so that users must pronounce another pronunciation of the polyphonic character to be able to recognize successfully when performing voice input. For example, the word "Xing" in the name Chen Lixing must be pronounced "plant '" to be successfully recognized, such as the user pronounced " '" can not be recognized correctly. For example, the word "乐" in the orchestra must be pronounced as " `" can be recognized, if it is pronounced as "ㄩ `" cannot be recognized correctly, and this voice input method is very different from the pronunciation habits of ordinary users

In addition, when a speech recognition system performs recognition, it usually uses the Viterbi Algorithm to calculate the probability value of the acoustic model corresponding to each word in the vocabulary for recognition, and such calculations are also the most computationally intensive for the speech recognition system. Therefore, if some of the same words are often repeatedly calculated, it will lead to unnecessary calculations of the system, and it will also cause a decrease in the recognition speed of the system, so we think about how to avoid repeated calculations of the same words to reduce the overall calculation.

Images