Research background

The purpose of the research is to develop novel algorithms for robust speech enhancement.

Traditional methods of speech enhancement apply filters with noise estimation therefore it is inevitable to have deterioration in original speech such as musical noise in severe SNR condition. This research will move away from the traditional approach to filter the noise, and reconstruct a clean speech signal from a set of acoustic features in noisy speech.


Research Team

Mr .Akihiro Kato, Dr. Ben Milner, Prof. Stephen Cox