Multiuser command speech recognition method and system and storage medium

Abstract

The invention discloses a multiuser command speech recognition method comprising the steps that at least one user speech instruction is acquired; existence of at least one wake-up word in at least oneuser speech instruction is recognized and detected according to the user voiceprint characteristics; existence of at least one executive command in at least one user speech instruction is recognizedand detected according to the user voiceprint characteristics; at least one wake-up word and at least one executive command are matched according to the user voiceprint characteristics so as to obtainat least one response instruction corresponding to at least one user; and at least one response instruction is matched and responded according to the response strategy. The wake-up words and the executive commands of multiple users are recognized through the voiceprint characteristics of the users and responded according to the response strategy. The wake-up words and the executive commands are separately recognized so that the speech recognition accuracy can be enhanced, the speech commands of the users are executed according to the response strategy and the speech commands can be clearly executed without confusion.

Description

technical field [0001] The invention relates to the field of speech recognition, in particular to a multi-person command speech recognition method, system and storage medium. Background technique [0002] Speech recognition technology refers to the technology by which a machine converts audio input into text or commands. In the speech recognition process, especially when the microphone array has a far-field recognition function, the sensitivity of audio signal acquisition will be relatively high. If there are other irrelevant people talking around the user, the microphone will collect the voices of multiple people at the same time, and the sound amplitude of the irrelevant part will be higher than the voice recognition threshold, resulting in the interaction between the user and the device and the interaction between other people around. Chatting is collected by the device, making it difficult for the speech recognition module to judge the user's intention and make a correc...

AI Technical Summary

Problems solved by technology

However, when many people speak in a noisy environment, the speech recognition effect cannot be improved.

[0004] With the rapid development of society and science and technology, human-computer interaction technology has become an important technology. Human-computer interaction applied to intelligent devices such as intelligent robots and smart phones requires speech recognition technology as the basis of human-computer interaction. In some human-computer interactions, individual commands are issued to smart devices. Voice recognition is also a single-person voice recognition. The recognition accuracy is not high, and the response on the smart device can only be responded by a single person. Only by recognizing the voice can it respond again, and it cannot respond continuously to the simultaneous commands of multiple users

Images