Setswana Speech Recognizer for Computer Based Applications
This study is the development and adaptation of Setswana speech recognizer into computer applications. Setswana database is used together with
sphinx decoder to build a generic Setswana speech recognizer. The recognizer is then adapted into a
developed game called General Knowledge Game (GKG) in which the user plays the game using Setswana speech. The developed recognizers level of accuracy was 60 percent on the worst case. The recognizer improves to more than 80 percent when shorter words are used to drive the application.
Participants have shown that they prefer using speech driven applications over traditional approach of using
mouse and keyboard. Analysis shows that though it is more effective to use keyboard and mouse to drive computer applications, users still prefer speech
interaction because HCI method is easy to learn particularly for the users who are semi-literate and illiterate. It shows that using traditional approach
(mouse and keyboard) requires some degree of literacy for someone to be competent while with speech interaction, anyone can use
Keywords: Setswana, Phoneme, Illiterate, General knowledge Game Database and Phonetic sequence
Download Full-Text
ABOUT THE AUTHORS
Oratile Leteane
Mr Oratile Leteane obtained B.Sc. degree majoring in Computer Science and Applied Mathematics at Nelson Mandela Metropolitan University (NMMU) in South Africa in 2007. Then, he obtained M.Sc. degree in Computer Science at University of Botswana (2012). Moreover, Leteane has worked as an assistant lecturer at ABM University College where he taught programming and systems analysis modules for one and half years. To date, Leteane is working as a systems analyst at Ministry of Education and skills development where his key responsibility is to support the development of business support systems for the ministry. Interests include research in development of advanced speech recognizing systems
Francis, J. Ogwu
Francis Joseph Ogwu has a B.Sc. Degree in Computer Science in 1981, M.Sc. Degree in Computer science Communications in 1984, a PhD in artificial intelligence Computer science in 1990 from Obafemi Awolowo University Nigeria. He worked as a lecturer in the above University for 18 years and he is presently working at the University of Botswana as a Professor of Computer Science. He has published with many journals in Europe, America, Asia and Africa. He has attended and chaired most conference round the world. He is a Special member of IASTED (International Association of Science and Technology for Development
Oratile Leteane
Mr Oratile Leteane obtained B.Sc. degree majoring in Computer Science and Applied Mathematics at Nelson Mandela Metropolitan University (NMMU) in South Africa in 2007. Then, he obtained M.Sc. degree in Computer Science at University of Botswana (2012). Moreover, Leteane has worked as an assistant lecturer at ABM University College where he taught programming and systems analysis modules for one and half years. To date, Leteane is working as a systems analyst at Ministry of Education and skills development where his key responsibility is to support the development of business support systems for the ministry. Interests include research in development of advanced speech recognizing systems
Francis, J. Ogwu
Francis Joseph Ogwu has a B.Sc. Degree in Computer Science in 1981, M.Sc. Degree in Computer science Communications in 1984, a PhD in artificial intelligence Computer science in 1990 from Obafemi Awolowo University Nigeria. He worked as a lecturer in the above University for 18 years and he is presently working at the University of Botswana as a Professor of Computer Science. He has published with many journals in Europe, America, Asia and Africa. He has attended and chaired most conference round the world. He is a Special member of IASTED (International Association of Science and Technology for Development