AUDIO SIGNAL TIME SEQUENCE PROCESSING METHOD, APPARATUS AND SYSTEM BASED ON NEURAL NETWORK, AND COMPUTER-READABLE STORAGE MEDIUM

摘要

An audio signal time sequence processing method and apparatus based on a neural network. The method comprises: creating a combined network model, wherein the combined network model comprises a first network and a second network (S1); acquiring a time-frequency pattern of an audio signal (S2); optimizing the time-frequency graph to obtain network input data (S3); using the network input data to train the first network, and performing feature extraction to obtain a multi-dimensional feature pattern (S4); using the multi-dimensional feature pattern to construct a new feature vector (S5); and inputting the new feature vector into the second network for training (S6). The method can solve the problem of an existing mapping transformation model based on a time sequence being unable to meet a multi-modal information application requirement.