Speech recognition method based on time domain convolution coding and decoding network

Abstract

The invention provides a speech recognition method based on a time domain convolution coding and decoding network, and the method comprises the steps: inputting audio information, obtaining audio features which comprise a Mel-frequency cepstrum coefficient and a linear prediction cepstrum coefficient, inputting the audio features into a pre-constructed neural network model to obtain a time sequence feature sequence, encoding and decoding the time sequence characteristic sequence through a time domain convolution network to obtain a group of output sequences, and obtaining a prediction result according to the output probability of each element in the output sequence and a preset mapping rule of the output sequence and a preset label sequence, the problem of speech recognition delay can be effectively solved.

Description

technical field [0001] The present invention relates to the field of speech recognition, in particular to a speech recognition method based on time-domain convolution codec network. Background technique [0002] At present, there are mainly traditional methods and deep learning methods in the field of speech recognition. Traditional methods mainly use HMM-based methods such as GMM-HMM or DNN-HMM to model each frame of speech; methods based on deep learning include convolutional neural networks, Deep neural networks, including recurrent neural networks, model large volumes of speech data. Disadvantages of existing technologies: Traditional HMM-based methods require frame-level labeling, which consumes a lot of time and manpower; the two-way recurrent network in deep learning methods cannot effectively solve the problem of recognition delay. Contents of the invention [0003] In view of the above existing problems in the prior art, the present invention proposes a speech re...

AI Technical Summary

Problems solved by technology

[0003] In view of the above existing problems in the prior art, the present invention proposes a speech recognition method based on time-domain convolution codec network, which mainly solves the problems of time-consuming labeling and delay in recognition in existing methods

Images