Skip to Main Content
Signal processing techniques are utilized in an attempt to discriminate speakers with laryngeal pathology from a normal population. These include classical periodicity analysis and discriminant analysis performed on a set of features representative of the associated all-pole LP models. These are applied to both the acoustic and electroglottographic waveforms. Preliminary results encourage further refinement of the technique and use of the laryngograph in this application.
Acoustics, Speech, and Signal Processing, IEEE International Conference on ICASSP '81. (Volume:6 )
Date of Conference: Apr 1981