Series feature extraction system and method aiming at general voice task in voice signal

Abstract

The invention discloses a series feature extraction system and method aiming at a general voice task in a voice signal. After an emotional corpus is built, voice preprocessing is performed; an audio file subjected to voice preprocessing is used as input to be connected with a voice feature extraction model; the voice feature extraction model can obtain basic information and prominent features of aspeaker through a voiceprint recognition layer; the basic information and the prominent features are fed into a voice recognition layer together; the information of the speaker and the text information are obtained; on the basis of extracting keywords, the information and low level descriptors are transmitted into an emotional recognition layer together; personalized voice features and emotionalfeatures are used for performing emotional recognition to obtain emotional relevant features; a voice feature extraction model can also obtain the basic information and the prominent features of the speaker through the voiceprint recognition layer; the basic information and the prominent features are fed into the voice recognition layer together with voice recognition features; the information ofthe speaker and the low level descriptors are transmitted into the emotional recognition layer together; and the emotional feature output is performed. The series feature extraction system and methodhave the advantages of great data scale, accurate emotion expression, high integration level and the like.

Description

technical field [0001] The invention relates to the field of signal processing and extraction, in particular to a feature extraction model of speech tasks. Background technique [0002] Speech is the most effective, natural and important form of communication for human beings. Realizing communication between humans and machines through voice requires machines to be intelligent enough to recognize human voices. With the development of machine learning, neural network and deep learning theory, the completion of speech recognition related tasks is gradually improving, which is of great help to the computer to understand the content of speech. Currently, speech recognition tasks mainly involve the following three recognition tasks: [0003] 1. Voiceprint recognition [0004] Voiceprint recognition, also known as speaker recognition, is a form of biometric recognition. It analyzes and processes the continuous speech signal of the speaker to extract discrete speech features, and...

AI Technical Summary

Problems solved by technology

[0009] However, there are many defects or deficiencies in the actual application or design of the above three recognition tasks

For example: Voiceprint recognition, speech recognition, and emotion recognition task models are not universal, input forms are not uniform, there is no universal solution, the integration accuracy is not high, the recognition accuracy rate of a single emotion recognition task is not high, etc.

Images