Audio classification model training and junk audio recognition method and device

Abstract

The invention discloses an audio classification model training and junk audio recognition method and device, and belongs to the technical field of artificial intelligence. The method comprises the steps: obtaining a plurality of audio sample segments and corresponding audio class tags, and constructing an audio classification model based on a convolutional neural network, a gating recursion unit and an attention mechanism; extracting an audio feature vector of each audio sample segment; based on each audio feature vector and the corresponding audio category label, training the audio classification model to obtain a trained audio classification model; receiving an identification request for a to-be-identified audio; segmenting the to-be-identified audio according to the identification request to obtain a plurality of audio segments; respectively identifying a first audio category of each audio segment by adopting the audio classification model; and obtaining the identification result ofthe to-be-identified audio based on the first audio category of each audio segment. According to the invention, the labor cost and the time cost are reduced, and the audio auditing efficiency is improved.

Description

Technical field [0001] This application relates to the field of artificial intelligence technology, and in particular to a method and device for training an audio classification model and identifying garbage audio. Background technique [0002] With the development of multimedia technology, the number and duration of audio have increased exponentially. When audio is uploaded to the Internet, it is usually necessary to review the audio to avoid pornography and Autonomous Sensory Meridian Response (ASMR) and other junk audio from spreading on the Internet. [0003] In traditional technology, manual review is usually used to determine whether the audio is junk audio after listening to the entire audio. Audio review efficiency is low, which consumes a lot of labor and time costs, and it is difficult to review mass audio. [0004] Therefore, there is a need for a technical solution for junk audio recognition that can reduce the labor cost and time cost and improve the efficiency of audio...

AI Technical Summary

Problems solved by technology

[0003] In traditional technology, manual review is usually used to judge whether the audio is junk audio after listening to the entire audio. The efficiency of audio review is low, which will consume a lot of labor and time costs, and it is difficult to review a large amount of audio.

Images