Speech input error correction method and system

Abstract

The invention relates to a voice input error correction method and system. The method includes converting an input voice into an identification text and an identification network; generating a lexical network according to the identification network; acquiring contents waiting to be corrected in the identification text; acquiring a modification-alternative replacement or rerecording instruction; if it is a modification-alternative replacement instruction, then generating a modification alternative corresponding to the contents waiting to be corrected according to the lexical network and selecting correct contents from the modification alternative to replace the contents waiting to be corrected; and if it is a rerecording instruction, then reacquiring a voice corresponding to the contents waiting to be corrected and correcting the contents waiting to be corrected. In the voice input error correction method and system, the modification alternative can be provided to modify error parts of the identification text and the error parts of the identification text can also be modified through rerecording part of the voice.

Description

technical field [0001] The invention belongs to the field of voice recognition, in particular to a voice input error correction method and system. Background technique [0002] Speech input is an input method that converts what people say into text through speech recognition. In practical applications, the environment faced by speech recognition is very complex, and speech content in various fields needs to be processed, and the recognition performance is difficult to achieve 100%. Therefore, for applications that require high text accuracy, the modification of the recognition result text It is particularly important. The current common error correction methods for voice input include: users edit text by themselves; users move the cursor to select text with errors to obtain candidates; The user feels a distinct sense of choppiness. [0003] Therefore, there is an urgent need for a convenient and smooth modification method to improve input efficiency and improve user experi...

AI Technical Summary

Problems solved by technology

In practical applications, the environment faced by speech recognition is very complex, and speech content in various fields needs to be processed, and the recognition performance is difficult to achieve 100%. Therefore, for applications that require high text accuracy, the modification of the recognition result text It is particularly important. The current common error correction methods for voice input include: users edit text by themselves; users move the cursor to select text with errors to obtain candidates; Users experience a noticeable sense of choppiness

Images