STUDY OF SOUND CLASSIFICATION USING DEEP LEARNING
Abstract
Sound plays a crucial part in every element of human life. Sound is a crucial component in the development of automated systems in a variety of domains, from personal security to essential monitoring. There are a few systems on the market now, but their efficiency is a worry for their use in real-world circumstances. Image classification and feature classification are the same as sound classification, just like other classification algorithms like machine learning. We also construct CNN architecture here. Deep learning architectures’ learning capabilities can be leveraged to construct sound categorization systems that overcome the inefficiency of standard methods. The goal of this paper is to use deep learning networks to classify environmental sounds based on the spectrograms that are created. The convolutional neural network (CNN) was trained using spectrogram images of environmental noises. For investigation, this paper used one dataset: Urbansound8K.There are 8732 sound clips (<=4s) of urban noises from ten classes in the collection. On this dataset, system was trained, and the accuracy acquired during training and testing was 98 percent and 91.92 percent respectively. The proposed approach for sound classification using spectrogram images of sounds can be efficiently employed to construct sound classification and recognition systems. Which can be used to distinguish audio evidence during crime investigation, remove noises and other useless sounds from music recording, or to classify different animals sounds in the forest.
Authors
Shalbin Benny , Arvind Kumar Tripathi , Rishabh Malik , Naman Mittal