Acoustic Sensor Network
The acoustic sensor networks research group is dedicated to a whole range of issues, from investigations into the interaction between the boundary conditions of the communication system and the requirements of acoustic signal processing to signal extraction and enhancement and the classification of acoustic events and scenes. The development of signal processing methods to protect privacy in order to prevent the unauthorised use of audio signals is also a key area of research.
Acoustic Event Detection
Sounds are an integral part of our environment. Be it a household or a working place, a restaurant or a busy street, a playground or a tram, sound events like 'footsteps', 'dog bark', 'cough', etc. carry immense information. Identifying these distinct events and their corresponding environment can therefore sometimes be very essential for different applications like audio surveillance. Though, the field of Automatic Event Detection (AED) is still in its infancy, numerous efforts are being made in the direction of detecting sound events and classifying the environments they are observed in. This research group focuses on developing techniques for robust and efficient detection of acoustic events from overlapping, real-life sound signals.