Audio scene recognition method based on integrated learning

Abstract

The invention discloses an audio scene identification method based on integrated learning. The method comprises the following steps: taking a single-channel audio signal, a left-channel audio signal,a right-channel audio signal and a central side-channel audio signal as three groups of training sets respectively; carrying out audio feature extraction on the three groups of training sets respectively for training three classifier networks respectively; taking the audio features of the training set as the input of a classifier network, training the classifier network, and identifying an audio scene in the existing test set according to the output result of the classifier network; and performing integrated learning on the three classifier networks, and identifying the audio scene according to the output after the integrated learning. Compared with the accuracy of a single classifier network, the accuracy of the method is averagely improved by 9.3%. The problem that the learning ability and the generalization ability of a single classifier network are insufficient is well solved, and comprehensive modeling can be carried out on complex audios in the whole data set. A high-performanceaudio scene recognition system can be obtained.

Description

technical field [0001] The invention relates to an audio scene recognition method. In particular, it relates to an audio scene recognition method based on ensemble learning for ensemble learning of multiple audio scene recognition sub-models. Background technique [0002] At present, the following methods are usually used for audio scene recognition. [0003] 1. Audio scene recognition description [0004] The data of audio scene recognition is directly collected in the real environment, so there must be overlapping sounds. Humans live in a complex audio environment and are very good at following specific sound sources while ignoring or simply acknowledging other sound sources. For example, we can have a conversation against a busy background consisting of other people talking or music. The performance of automatic classification for audio scene recognition has been greatly limited in this task. Acoustic mixture signals contain multiple simultaneous sound events, and th...

AI Technical Summary

Problems solved by technology

In a multi-source environment, the performance of such a system is very limited

Images