Speech emotion recognition method and system based on semi-supervised adversarial variation self-coding

Abstract

The invention discloses a speech emotion recognition method and system based on semi-supervised adversarial variation self-encoding, and the method comprises the steps: S1, constructing a generative adversarial network, and constructing a speech emotion recognition model through the combination of a semi-supervised variation self-encoding model and the generative adversarial network, wherein data with emotion labels in input data and corresponding emotion labels are used as input, data without emotion labels in the input data are used as emotion label attribute missing types for processing, feature probability distribution of the input data in a hidden layer is learned through the generative adversarial network, and an SSAVAE model is constructed; S2, training the constructed SSAVAE model by using a training set; and S3, inputting to-be-processed speech emotion data, and inputting the to-be-processed speech emotion data into the trained SSAVAE model to obtain an emotion recognition result. The method has the advantages of being simple in implementation method, high in recognition precision, good in generalization ability and data disturbance resistance and the like.

Description

technical field [0001] The invention relates to the technical field of speech emotion recognition, in particular to a method and system for speech emotion recognition based on semi-supervised confrontational variational self-encoding. Background technique [0002] Speech emotion recognition aims to extract emotion-related features from speech signals, identify the emotional state of the current speaker, and enhance the naturalness of human-computer interaction. It can be widely used in human-computer interaction, voice customer service, vehicle-mounted systems, etc. Scenes. Speech emotion recognition is one of the tasks belonging to pattern recognition. Different supervised learning models can be used to construct speech emotion recognition systems with good recognition performance, such as hidden Markov models, Gaussian mixture models, support vector machines, etc. However, the above models are all shallow model structures, which limit model learning Deep Emotional Featur...

AI Technical Summary

Problems solved by technology

[0005](1) Since we usually only focus on learning the common representation of labeled data and unlabeled data in the feature space, the learned common representation is the low-dimensional feature of the input data Mapping, so the generalization ability of the above semi-supervised learning method is weak, and its performance is easily affected by the disturbance of the input data

[0006] (2) The quality of emotional feature representation will directly affect the recognition performance of the model. The above-mentioned model constructed using the semi-supervised learning method cannot fully represent emotional features. Representation quality still needs to be improved, affecting the accuracy of speech emotion recognition

Images