Cloud interaction voice interaction method

Abstract

The invention provides a cloud interaction voice interaction method, which is characterized in that on the basis of communication connection between machine terminal equipment and cloud equipment, strong calculation capability and perfect algorithm model design of the cloud equipment per se are used; voice signals received by the machine terminal equipment are subjected to automatic echo elimination processing and detection recognition processing relevant to preset wake-up words through the cloud equipment. The voice interaction method can be used for eliminating echo signals more comprehensively and efficiently as well as detecting and recognizing the preset wake-up words in the voice signals more precisely and fast, so that the occurrence of conditions of wake-up by mistake or self-excitation wake-up due to unthorough elimination of the echo signals of the machine terminal equipment can be effectively avoided.

Description

technical field [0001] The invention relates to the technical field of voice recognition, in particular to a voice interaction method for cloud interaction. Background technique [0002] With the development of artificial intelligence technology, it is widely used in people's life and work, and people can solve problems in specific fields in a more convenient way through artificial intelligence technology. The reason why artificial intelligence technology can be integrated into people's life and work is mainly due to the continuous improvement of human-computer interaction technology. Human-computer interaction technology enables people to easily and conveniently realize the communication and interaction between humans and machines by using the common communication methods between humans. At present, the main implementation methods of human-computer interaction technology include human-computer voice interaction or human-computer action interaction; among them, the human-co...

AI Technical Summary

Problems solved by technology

Since the original sound signal received by the machine terminal device generally has a large number of echo signals, interference signals and sound signals related to non-wake-up words, in order to realize the accurate wake-up operation of the machine terminal device, this requires the machine terminal device to perform high-precision automatic Echo cancellation and wake-up word detection and recognition calculation processing, and these calculation processing put forward high requirements on the computing power of the machine terminal equipment itself, but the existing machine terminal equipment has different levels of local computing power, algorithms and model design Due to the limitation of the computer, its computing and processing performance cannot meet the corresponding accuracy requirements for voice interaction recognition.

Images