Voice detection method and device, electronic equipment and storage medium

Abstract

The invention discloses a voice detection method and device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: inputting to-be-detected audio information into a voice classification model to obtain a first result, and enabling the first result to be used for determining a corresponding relation between theaudio information and a target object, wherein the voice classification model is obtained by training audio training information, and the voice classification model at least comprises a convolutionalneural network, a recurrent neural network and an attention mechanism model which are cascaded in sequence, and a fusion layer used for fusing output results of the convolutional neural network and the attention mechanism model. Thus, the audio information is detected based on the trained voice classification model to determine the corresponding relationship between the audio information and the target object, manual auditing is not needed, the labor cost and the time cost are reduced, the audio auditing efficiency is improved, and auditing and filtering of large-scale voice data can be realized.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, and more specifically, relates to a voice detection method, device, electronic equipment and storage medium. Background technique [0002] In the information age of multimedia explosion, the number and duration of voices are increasing exponentially. When the voice is uploaded to the network, it usually needs to be audited to avoid the spread of spam voice in the network. In some scenarios, the voice of a certain group of people is garbage voice. [0003] At present, in order to identify the above-mentioned spam voices, manual review is usually used, that is, to judge whether a piece of voice contains the voice of a specific group of people by manually listening to the voice. This requires the reviewer to listen to the entire speech. In theory, a reviewer can only review the speech of no more than 24 hours a day, which is inefficient and requires high labor and time co...

AI Technical Summary

Problems solved by technology

This requires the reviewer to listen to the entire speech. In theory, a reviewer can only review the speech for no more than 24 hours a day, which is inefficient and requires high labor and time costs.

Moreover, the manual review method cannot complete the review and filtering of large-scale voice data

Images