Sixth European Conference on Speech Communication and Technology
(EUROSPEECH'99)

Budapest, Hungary
September 5-9, 1999

A Study of Speech Recognition for the Elderly

Fumihiro Yato (1), Naomi Inoue (2), Kazuo Hashimoto (2)

(1) Teikyo University of Science & Technology, Uenohara-machi, Yamanashi, Japan
(2) KDD Research & Development Laboratories, Kamifukuoka-shi, Saitama, Japan

The elderly have needs for voice operated interfaces to manipulate various systems. However, only few results were reported about the performance of the current speech recognition technology with the elderly. In this paper we report our speech recognition experiments for 469 Japanese elderly persons belong to the age range 59 to 85. The speech recognition is executed on PC by using IBM Via Voice. The utterances of 600 place names are used for our speech recognition experiments. The average error rate becomes about 7 %, which seems to be a little bit lower than that of young adults. The paper presents the detailed analysis of the results of our experiments.


Full Paper (PDF)   Gnu-Zipped Postscript

Bibliographic reference.  Yato, Fumihiro / Inoue, Naomi / Hashimoto, Kazuo (1999): "A study of speech recognition for the elderly", In EUROSPEECH'99, 101-104.