Speaker Identification
Fourth year undergraduate student, Brett Squires, applyied machine learning to the problem of identifying a person from the characteristics of a sentence spoken into a microphone. This was successfully demonstrated and later work was applied to a security system based on this method. Andrew Taylor also taken up this approach to identify bird, frog and whale species for environmental research.
Sound Source Localisation and Characterisation
Ben Rudzyn followed on from Brett’s work to develop a system for localising sounds using a microphone array. The system also incorporated learning to recognise the type of sound source.
Publications
Squires, B. and Sammut, C. (1995). Automatic Speaker Recognition: An Application for Machine learning. In A. Predietis & S. Russell (Eds.), Twelfth International Conference on Machine Learning. Morgan Kaufmann.
Rudzyn, B., Kadous, W. and Sammut, C. (2007). Real time robot audition system incorporating both 3D sound source localisation and voice characterisation. In International Conference on Robotics and Automation, Rome.