Speech recognition method and system, computer system and computer readable storage medium

Abstract

The disclosure provides a speech recognition method. The method comprises the following steps: acquiring a text file obtained by preprocessing a speech file; converting the text file according to a predetermined pinyin coding rule to obtain a corresponding first conversion code file; matching the first conversion code file with a second conversion code file to obtain a matching result, wherein thesecond conversion code file is obtained by converting each keyword in a keyword text set according to the predetermined pinyin coding rule; and recognizing the speech file according to the matching result. The disclosure also provides a speech recognition system, a computer system and a computer readable storage medium.

Description

technical field [0001] The present disclosure relates to the technical field of speech recognition, and more specifically, to a speech recognition method, system, computer system, and computer-readable storage medium. Background technique [0002] At present, speech recognition basically adopts the recognition method of first recognizing the speech file as a text file, and then performing keyword matching on the text file. If the corresponding keyword can be matched in the text file, it is determined that the corresponding speech file contains the keyword. However, since the noise and slight dialect contained in the voice file will be recognized as other words, for example, in the customer service call, the customer service staff generally speaks Mandarin, while many users speak dialects. In this case, it is easy to As a result, the corresponding keyword cannot be found in the subsequent keyword matching process. [0003] However, in the process of realizing the concept of...

AI Technical Summary

Problems solved by technology

However, because the noise and slight dialect contained in the voice file will be recognized as other words, for example, in the customer service call, the customer service staff generally speaks Mandarin, while many users speak dialects. In this case, it is easy to As a result, the corresponding keywords cannot be found in the subsequent keyword matching process

[0003] However, in the process of realizing the concept of the present disclosure, the inventors have found that there are at least the following problems in the related technologies: the above-mentioned situation is relatively common at present, because the speech training set based on the speech recognition system is generally standard Mandarin and the speech rate is appropriate, without Speech environment affected by noise, and the speech files that use this speech recognition system for speech recognition are generally not standard Mandarin or even dialects, and the speed of speech varies greatly, and the speech environment affected by noise will lead to speech recognition failure. very low accuracy

Images