________________________________________ † : Corresponding Author Input Device of Note Taking System for Hearing Impaired Student Shuichi Seto † Kinjo College, 1200 Kasama-machi, Hakusan, Ishikawa, 924-8511 JAPAN Email: seto@kinjo.ac.jp Kimikazu Sugimori 1 , Yuko Shimomura 2 , Hiroyuki Kawabe 3 and Tsuyoshi Kimura 4 Kinjo University, 1200 Kasama-machi, Hakusan, Ishikawa, 924-8511 JAPAN Email: sugimori@kinjo.ac.jp 1 shimo@kinjo.ac.jp 2 kawabe@kinjo.ac.jp 3 kimura-t@kinjo.ac.jp 4 Abstract. Generally, attendance at lecture of hearing impaired is supported by students in the same university. Most of who are using support systems based upon personal computer. However, it is very difficult for the supporting staffs to simultaneously take a note while hearing a lecture, and to copy contents written on a blackboard. Therefore, it is to be desired that exports of note taking support the hearing impaired. The purpose of this research project is to take a lecture note without delay by many beginners of the note taking and a text mining technology, which is an alternative way of a skilled note taker. For our system, the text data of lecture note are input by many beginners of note taking. Therefore it is important to choice a suitable device. For our purpose, we have three types of candidate of input devices, a keyboard, a handwriting input device and a voice recognition system. The keyboard is the most popular input device for PC users; however, it is necessary to train users in order to type correctly and fast. The handwriting input device is easy to use for beginners, but late. The voice recognition system which converts voice into letters is easy to use, but it is necessary to tune the system for an operator. In this study, we examine above three types of devices and report the result. Keywords: Input device, Note-taking, Hearing impaired student. 1. INTRODUCTION Recently impaired students in University increase. Admitting them in University, we must prepare some kind of supports for them in order them to understand teacher’s speeches in lecture. Candidates for method to support are, for example, a sign-language, a note-taking, and a condensed transcription. The second and the third methods are transcriptions made from speed. It is hard to gain these skills within a short time. For example, it takes four months or more to gain fundamental skills of the condensed transcription and about two years to be active as a volunteer. As a result, we have a problem that there are few volunteers of the condensed transcription (Hatakeyama et al. 1994, PEPNet-Japan 2006). Though courses for training the condensed transcription as a counter measure are held, we don’t have many active volunteers. Consequently, some systems that a few volunteers are made the best use of though a computer network, for example, one that a skilled expert inputs by voice into a voice recognition system connected though the Internet (Kobayashi 2004), and, one that the voice of teacher is transmitted to a skilled volunteer by the Internet and the volunteer sends the condensed transcription back to a hearing impaired student in the class room (Tachiiri et al. 2003), are proposed. Both systems require the expert of voice input or skilled note-taker. The objective of this research is to construct a note- taking system in which many unskilled volunteers enrolled at university are accepted as alternatives to a skilled volunteer. On usual note-taking systems for the hearing impaired students, skilled note-takers take a key role, but any systems in which many unskilled beginners support the hearing impaired have not been reported. Moreover, we have not known about systems in which contents of lecture are gathered and are delivered to the hearing impaired within real time with the help of a data mining technique. In this report, first we give a concert of our note-taking system, next we explain about features of some input devices, experiments of text-input of speech and results of 2135