• Ruturaj Ghodke
• Prajan Shukla
• Vedita Kharabe

In life, sound plays a very vital role. Just like sight, humans perceive sound and gain crucial information from it. Even without sight, sounds help us to understand what is happening around us. Today, many security systems like cameras or trackers are based on either sight or motion. Even for the disabled, visual aid is provided, and hence the importance of sound here is neglected. Environmental sounds when analyzed and identified, give us an idea about the presence of things around us. Urban environmental sounds are so unstructured in nature that it is not easy to classify them, but they contain patterns of temporal spectrum that assist in analysis and classification. Using these techniques, a machine can learn to identify sounds and accurately identify them. Hence, we can also use the trained system for surveillance in cameras where the sight is out of range, to aid people with difficulty in hearing and many other applications. The aim of this research is to identify and classify a given environmental sound by using a CNN for classification. We feed an environmental sound audio file to the system after converting it to a spectrogram (visual representation) of the sound and predict the output of the type of sound fed in. Since spectrograms are nothing but images, we use CNN as it works best with image classification and computer vision.

Deep Learning, Audio Processing, Spectrogram, CNN.

"Urban Sound Classification using Deep Learning", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.10, Issue 8, page no.e272-e279, August-2023, Available :http://www.jetir.org/papers/JETIR2308432.pdf

"Urban Sound Classification using Deep Learning", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.10, Issue 8, page no. ppe272-e279, August-2023, Available at : http://www.jetir.org/papers/JETIR2308432.pdf