An Analysis of Audio Classification Techniques using Deep Learning Architectures

Failure to classify audio data with high efficiency causes major setbacks in audio processing, voice recognition, and noise cancellation. To find the best possible neural network models for audio classification, this article shows the steps in the experiments performed in the newly designed CF Model and CF Clean Model in both CNN and RNN and compares the results with some existing models such as DCNN and PiczakCNN. To get a clear view of the consistency of the results, three different datasets have been experimented on, which are UrbanSound8k, FSDKaggle2018 and ESC-50. It also performs best in terms of the training and testing dataset based on accuracy and loss percentage. Finally, this article shows influences about the envelope function, normalization, segmentation, regularization techniques, and dropout layers in the overall progress.