Automatic selection of recognition errors by respeaking the intended text.

Keith Vertanen, Per Ola Kristensson

Browse the full ASRU paper archive.