The goal of this thesis is the development of a software prototype that can annotate dangerous sound events. The thesis includes building a pipeline that can detect special such events and developing a Deep Neural Network (DNN) that can classify sound events as dangerous. Furthermore, the pipeline will provide a table-like file (e.g. CSV) which contains for each audio-file an entry with the filename, the probability, the classification and the timestamps (start and end) of the occurred sound-event. The software and thesis are developed in collaboration with MED-EL.
Supervisor: Michael Felderer