Voice classification method and device, server and storage medium

Abstract

The embodiment of the invention discloses a voice classification method and device, a server and a storage medium, wherein the voice classification method comprises: acquiring the MFCC feature matrixof a target short voice by using a Mel frequency cepstrum coefficient MFCC algorithm, and converting the MFCC feature matrix into a target image; based on a deep learning model, extracting the targetimage feature of the target image; inputting the target image feature into a pre-trained voice classifier, and outputting the category of the target short voice. The embodiment of the invention solvesthe problem that the existing voice classification method ignores the deep information of the voice content, can only roughly evaluate the voices with a large content difference, and realizes an effect of classifying and processing the voice data quickly and effectively.

Description

technical field

[0001] The embodiment of the present invention relates to the application field of Internet technology, and in particular to a voice classification method, device, server and storage medium. Background technique

[0002] With the rapid development of the Internet industry and the expansion of voice information, how to quickly and accurately classify voice data in massive information and save computing resources is a difficult point at present.

[0003] The existing speech classification method usually calculates the MFCC features of each frame in the speech data, and then stitches the MFCC features of each frame into the overall features of short speech, trains a classifier and performs feature classification, and then obtains classification labels. However, based on the general speech classification method, the deep information of the speech content is ignored, and only a rough assessment can be made on the speech with a large difference in content. Conten...

Images