Speech class recognition method, apparatus, computer device, and storage medium

Abstract

The present application relates to the field of speech recognition, and provides a method, apparatus, computer device and storage medium for recognizing speech categories. The method comprises: acquiring first speech information to be recognized and converting the first speech information into a first language spectrum map; inputting the first language spectrum map into a preset speech classification model to obtain a classification result of the first language spectrum map, and taking the classification result as a class of the first speech information; wherein, the speech classification model is trained based on a depth convolution neural network by using a language map of a known emotion class or a character class; The speech recognition method, apparatus, computer device and storage medium provided in the present application are convenient to improve the effect of emotion and character classification in speech information.

Description

technical field [0001] The present application relates to the technical field of speech recognition, and in particular to a speech category recognition method, device, computer equipment and storage medium. Background technique [0002] At present, the focus of speech emotion and personality recognition is mainly on the extraction of acoustic features. In the existing speech emotion recognition technology, the acoustic features used for recognition include prosodic features, voice quality features, spectral correlation features, and fusion features formed by filtering the above features. These features are mainly concentrated in the time domain or frequency domain alone. For speech signals with associated changes in the time domain and frequency domain features, the above features often lose part of the feature information, which in turn affects the effect of emotion and personality recognition; and the above acoustic features In the extraction process, it will be affected ...

AI Technical Summary

Problems solved by technology

These features are mainly concentrated in the time domain or frequency domain alone. For speech signals with associated changes in the time domain and frequency domain features, the above features often lose part of the feature information, which in turn affects the effect of emotion and personality recognition; and the above acoustic features In the extraction process, it will be affected by some factors that have nothing to do with emotion and personality (such as speech content, speaker, environment, etc.)

These irrelevant factors are included in the extracted acoustic features, which will also greatly affect the effect of emotion and personality classification

Images