C. Stephanidis (Ed.): Posters, Part I, HCII 2013, CCIS 373, pp. 286–290, 2013.
© Springer-Verlag Berlin Heidelberg 2013
Development of Recognition System of Japanese Sign
Language Using 3D Image Sensor
Yanhua Sun, Noriaki Kuwahara, and Kazunari Morimoto
Graduate School of Science and Technology Kyoto Institute of Technology Matsugasaki,
Sakyo, Kyoto 606-85858, Japan
oo_alison@hotmail.com
{noriaki.kuwahara,morix119}@gmail.com
Abstract. The population of Japanese people with disabilities is growing day
by day. And the population of sign language translator is too few to support
them. In general life, people communicate with others through conversation but
this is obviously impossible for deaf mute people who use sign language to
communicate. So it is necessary to explore a recognition system of sign
language which can help the deaf and mute to keep in touch with others. In
order to solve that problem, a large amount of researches related to recognition
system development and establishment have been reported by previous
literatures. However, current paper introduced a novel method for system
developing. In this paper, 3D sensors called Kinect were employed for hand
gesture dataset’s collection following by data dealing from transformation
matrix based on specific formulas. Although the hand gesture can be captured,
but there still are a lot of noises left, so PCL (Point Cloud Library) was applied
to do the 3D data processing.
Keywords: Japanese Sign Language, Kinect, Recognition of JSL, 3D Sensor.
1 Introduction
In Japan, the population of people with disabilities who are deaf and mute or hard of
hearing and speaking is about 2.7million. It is said that about 3 thousand people do a
job as a sign language translator. There are so little people to support the deaf and
mute. In general life, people take the communication with the others through talking,
but it is impossible for the deaf mute people, who can take the communication with
others just using sign language. So it is necessary to explore a recognition system of
sign language, which can help the deaf and mute to keep touch with others.
So developing a recognition system of sign language is the purpose of this paper.
About the system, Sign language recognition is such a system which translate the
picture of the sign language to data, and then translate the data to words, sentences or
sounds. This paper presents the front part of the recognition system. Several ways are
used to research the sign language recognition, such as Facial movement analysis in
ASL [5], human detection [3], 3D model tracking [4] and so on. But in most of the
researches, even though they carry important grammatical and prosodic information,
the rate of the recognition is still low.