A cross-database speech emotion recognition method based on deep domain adaptive convolutional neural network

Abstract

The invention discloses a cross-database speech emotion recognition method based on a deep domain adaptive convolutional neural network, comprising: (1) acquiring training databases and testing databases with different languages ​​(2) converting speech signals in the training database and testing database to Respectively process and obtain the spectrogram of each segment of the speech signal; (3) set up a convolutional neural network; (4) input the speech signal spectrogram of the training database and the test database into the convolutional neural network respectively for training. During training, first calculate the training database The maximum mean difference between the output of the fully connected layer corresponding to the speech signal spectrogram of the test database, and then calculate the cross entropy between the softmax layer output of the training database and its emotional category label, and finally add the maximum mean difference and cross entropy as The network loss uses the backpropagation algorithm to update the network parameters to complete the network training; (5) Obtain the spectrum map of the speech signal to be recognized, input the trained deep convolutional neural network, and output the emotional category. The accuracy rate of the present invention is higher.

Description

technical field [0001] The invention relates to speech data emotion recognition, in particular to a method for cross-database speech emotion recognition based on deep domain adaptive convolutional neural network. Background technique [0002] Speech emotion recognition is a research hotspot in the field of pattern recognition and artificial intelligence, and has broad application prospects. Traditional voice emotion recognition is often trained and tested on a single voice database, but in real life, the voice data of the training set and the test set are often very different, such as from different languages, so it is performed on different voice databases. Speech emotion recognition is closer to real life scenarios, which is a cross-database speech emotion recognition problem. The difficulty of cross-database speech emotion recognition is to extract appropriate speech features and reduce the difference in feature distribution between source database data and target databa...

AI Technical Summary

Problems solved by technology

The difficulty of cross-database speech emotion recognition is to extract appropriate speech features and reduce the difference in feature distribution between source database data and target database data

Images