Suche

Rare Acoustic Event Detection


Description While sound event classification has been an inherent part of audio processing research, rare sound event detection is relativity new task. A rare sound is such that occurs infrequently, is unexpected, and can be a cause of alarm.
Task In this thesis, the student(s) will be analysing a range of convolutional and recurrent neural networks for detecting rare acoustic events as well as develop their own system.
Utilises Convolutional Recurrent Neural Networks, auDeep, Deep Spectrum System.
Requirements Preliminary knowledge in Machine Learning, Good programming skills (e.g. Python, C++).
Languages English and German.
Supervisor Shahin Amiriparian, M. Sc. (shahin.amiriparian@informatik.uni-augsburg.de)