Voice awakening method through gesture recognition, smart TV and storage medium

Abstract

The invention discloses a voice awakening method through gesture recognition, a smart TV and a storage medium. The method comprises the following steps: arranging a camera for acquiring a gesture operation on the smart TV in advance, and enabling a predetermined operation gesture acquired by the camera to correspond to an operation instruction of the smart TV, and enabling the predetermined operation gesture to correspond to an operation instruction for calling voice recognition; monitoring image information through the camera of the smart TV, and calling a voice recognition module when the camera captures the predetermined operation gesture corresponding to the n operation instruction for calling voice recognition; controlling to turn on a microphone through the voice recognition module so as to acquire a corresponding voice command, and controlling the smart TV to execute the corresponding operation. Through the method disclosed by the invention, the technical short board of the voice awakening can be effectively solved to provide convenience for the user, and the technical realization cost can be effectively lowered to provide convenience for the user.

Description

technical field [0001] The invention relates to the technical field of smart TVs, in particular to a voice wake-up method through gesture recognition, a smart TV and a storage medium. Background technique [0002] The current smart TVs have popularized speech recognition applications, and operations such as query and control can be performed through voice. The introduction of far-field voice technology makes the human-computer interaction of the TV completely separated from the remote control, and directly wakes up the background engine through voice to achieve the effect of human-computer intelligent dialogue. [0003] The defect of the prior art solution is that the algorithm needs to eliminate the environmental noise and the sound of the TV itself, otherwise the sound of the TV interferes with the sound pickup effect of the target sound source. The defect of the smart TV algorithm in the prior art is that when the volume of the TV exceeds a certain limit, far exceeding t...

AI Technical Summary

Problems solved by technology

[0005] In actual use, the volume of a normal person’s speech is about 65dB, and the maximum volume of a TV can reach more than 140 dB. When the volume of a TV is much higher than the voice of a person, generally exceeding 20dB, it is difficult for the microphone to pick up sound effectively. The sound enhancement algorithm has been optimized, but it still cannot meet the needs of voice wake-up under high volume conditions

Therefore, under the existing technical conditions, there is a bottleneck of voice wake-up technology

Images