Improved DCNN music genre classification method

Abstract

The invention discloses an improved DCNN music genre classification method. The method comprises the steps of inputting a training set and a verification set, extracting MFCC features of the audio information, generating a frequency spectrum, carrying out frequency spectrum cutting, inputting a network model, training a model, verifying the model, judging whether a specified batch is reached or not, and outputting a model. According to the method, self-adaption of channel dimensions is achieved through a function, so that the coverage range of interaction of local area cross channels is ensured, an ECA module is more effectively integrated into an existing DCNN architecture, obvious performance gain is brought to a network model, and then the working efficiency of music genre classification is improved. Through the Mel-frequency cepstrum coefficient, the perception characteristics of a human auditory system are simulated, and the classification precision is further improved.

Description

technical field [0001] The invention belongs to the technical field of hyperspectral image classification, and in particular relates to an improved DCNN music genre classification method. Background technique [0002] At this stage, traditional music genre classification methods have been slowly replaced by deep learning methods. Compared with traditional methods, the method of genre classification through feature learning and deep structure mainly has the following advantages: (1) avoiding too professional music theory knowledge and technology, and does not require users to design artificial features; (2) has good end-to-end learning It has certain advantages in solving problems related to music genres; (3) It can greatly reduce the burden on professionals and improve related work efficiency. [0003] Deep Convolutional Neural Network (DCNN), as one of the most widely used deep learning models in the field of music genre classification, has also achieved good application r...

AI Technical Summary

Problems solved by technology

However, while improving the classification method and optimizing the classification effect, it also increases the computational burden of the model structure accordingly.

In the feature representation of music data, Liu Shenglan et al. obtained the corresponding attention probability distribution through a parallel attention model, and aggregated features through maximum pooling. Although stronger learning ability can be achieved, the complexity of the model is also high. increase accordingly

Images