The University of Oslo is closed. The PhD defence and trial lecture will therefore be fully digital and streamed directly using Zoom. The host of the session will moderate the technicalities while the chair of the defence will moderate the disputation.
Ex auditorio questions: the chair of the defence will invite the audience to ask ex auditorio questions either written or oral. This can be requested by clicking 'Participants -> Raise hand'.
Trial lecture
"Filter design based on convex optimization"
Main research findings
Microphone arrays constitute an increasingly important component in modern information systems. This thesis deals with various ways of overcoming the limitations of beamforming with microphone arrays, mainly the limited resolving power and ambiguities caused by sidelobes. We have developed a framework for camera-assisted beamforming and used this to detect speech activity of multiple simultaneous speakers, which is a requirement for e.g. robust remote voice interfaces.
We have investigated the problem of deblurring of acoustic images and proposed a family of computationally efficient algorithms applicable to a small number of sources. We have shown through large-scale simulations and experiments that the proposed methods outperform traditional methods in certain aspects.
Finally, we have proposed a variation on a beamforming method for concentric circular arrays with partially frequency-invariant response and shown how our proposal mitigates the problems normally associated with this processing method.
Contact information to Department: Mozhdeh Sheibani Harat