Comparing speaker-dependent and speaker-adaptive acoustic models for recognizing dysarthric speech

Frank Rudzicz
2007 Proceedings of the 9th international ACM SIGACCESS conference on Computers and accessibility - Assets '07  
Data • We use the annotated Nemours database [1]. • This contains 11 dysarthric male speakers, each producing 74 nonsense sentences of the form The (N 0 ) is (V)ing the (N 1 ).
doi:10.1145/1296843.1296899 dblp:conf/assets/Rudzicz07 fatcat:zoxujyjx5ndhxpus7oxim5l7qi