Improved acoustic modeling for automatic dysarthric speech recognition | IEEE Conference Publication | IEEE Xplore