Article ID Journal Published Year Pages File Type
10368526 Computer Speech & Language 2014 16 Pages PDF
Abstract
► We develop an automatic running speech intelligibility assessment method. ► We derive new text-independent speaker features from a phonological representation. ► We validate the method on a new corpus of patients treated for head and neck cancer. ► We show that it is as reliable as a human listener and able to track a speaker's progress.
Related Topics
Physical Sciences and Engineering Computer Science Signal Processing
Authors
, , , ,