Large-scale audio dataset for emergency vehicle sirens and road noises

This section discusses the methodology used to collect sound data and extract its features. This research is performed at National Centre for Big data and Cloud Computing, Ziauddin University (NCBC-ZU), Karachi, Pakistan. Sound data is collected using various techniques to build a dataset: (1) Data collection through mic integrated HD Camera, (2) Online Sources, and (3) Placing Emergency Siren manually.

Data collection procedures

Various methods have been implemented to build a database. Firstly, a microphone mounted on the HD camera is used to capture the sound of an ambulance on the roads. NCBC-ZU lab has installed microphone-integrated HD cameras at twenty different locations in Karachi. However, only four locations have been selected for data collection. The first location is a road…

Continue Reading

News Source: