The voice controlled Daisy Reader and beyond
With the Daisy Reader App visually impaired people can listen at any time to talking books, newspapers and magazines. The Dutch Library Service for the blind and their production partner Dedicon notice an increase every year in the number of users of the Daisy Reader App and the number of titles streamed (in 2015 more than 400,000 streams).
The app is accessible, e.g. through the VoiceOver or Talkback functionality of mobile devices (tap and swipe), but it is not always easy to learn this way of operating, especially for older visually impaired people. They are less familiar with smartphone or tablet and cannot use the Daisy Reader App (correctly), let alone control it by means of tap and swipe. For this target group voice control using automatic speech recognition (ASR) provides a solution. This technology is growing rapidly and is seen as a welcome addition to the technological potential. Dedicon is leading a project to realize voice control for the Daisy App with the speech technology experts of CLST at Radboud University.
Voice control can also be applied for increasing the accessibility in other interfaces, for example about navigating the websites, and searching and ordering in webshops. Even further beyond, voice control for visually impaired will be available to control house equipment such as refrigerators and coffee machines, or to communicate with (health care) robots or virtual assistants.
Dr. Henk van den Heuvel
Dr. H. van den Heuvel studied German Language and Literature (main topic phonetics) in Utrecht. In 1996 he defended his PhD thesis entitled ’Speaker variability in acoustic properties of Dutch phoneme realisations’. He co-ordinated the work on orthographic transcriptions in various projects for KPN, Philips, Temic, and CGN. Since 2003 he is director of the Centre for Language and Speech Technology (CLST) at the Radboud University in Nijmegen. For CLST he was involved in various projects on resources and technology that are needed to improve automatic speech recognition and automatic name recognition.
Dr. Emre Yilmaz
Dr. Emre Yilmaz was born on 15 September 1986 in Konya, Turkey. He received the B.Sc. degree in electrical and electronics engineering from the Middle East Technical University (METU), Turkey in 2008 and the M.Sc. degree in electrical engineering from the Royal Institute of Technology (KTH), Sweden in 2010. Then, he worked as a part-time researcher in the Institute of Communication Systems and Data Processing (IND), RWTH Aachen, Germany. He joined the Department of Electrical Engineering (ESAT), KU Leuven, Belgium as a Ph.D. candidate in January, 2011 and received the Ph.D. degree in May, 2015. He is currently working as a postdoctoral researcher at the CLST group in Radboud University Nijmegen, Netherlands. His research interests are noise robust automatic speech recognition (ASR), ASR of low-resourced languages, ASR of children and dysarthric speech and medical applications of ASR.