Data Augmentation Using Generative Adversarial Network for Environmental Sound Classification

Various types of deep learning architecture have been steadily gaining impetus for automatic environmental sound classification. However, the relative paucity of publicly accessible dataset hinders any further improvement in this direction. This work has two principal contributions. First, we put forward a deep learning framework employing convolutional neural network for automatic environmental sound classification. Second, we investigate the possibility of generating synthetic data using data augmentation. We suggest a novel technique for audio data augmentation using a generative adversarial network (GAN). The proposed model along with data augmentation is assessed on the UrbanSound8K dataset. The results authenticate that the suggested method surpasses state-of-the-art methods for data augmentation.