Detection of Acoustic Scenes and Events using Audio Analysis – A Survey

Abstract

The field of Acoustic Event Detection is generally overshadowed by the traditional field of Automatic Speech Recognition. But our auditory environment is made up of many sounds other than speech. Most of the current surveillance systems, diagnostic systems, etc. use video cameras and other sensors. But acoustic sensors are not used frequently because of lack of event detection accuracy in case of overlapped events. This motivates the need for developing a system that is able to identify acoustic scenes and event correctly. Researchers have worked on the acoustic events of diverse fields such as bird species recognition, classification of different type of vehicles, etc. This work has considered the all such acoustic scenes in which humans interact with the environment directly or indirectly. In this work, we have provided an overview of major studies done in this field along with the suggestions regarding what can be done in future in this field. We have also studied the various types of features related to the various acoustic scenes and events. We have summarized the recent open-source datasets and challenges conducted by IEEE AASP done in this field, which could help the future researcher to develop a foundation in this field.