Development of Web-Based Remote Speech-to-Text Interpretation System captiOnline

抄録

For ensuring information accessibility for hearing impaired, speech-to-text interpretation captioning is used. Usually, a text interpreter at the site interprets the speech of a speaker and enters a text into a PC. The text is then displayed through a projector or other devices to the users. For Japanese typing, in 1998, a method was developed for a pair of interpreters to collaboratively enter a sentence in which they used dedicated software to convert speech to text and present the resulting text. In the recent years, remote speech-to-text interpretation using the internet has been studied to overcome place and time restrictions of interpreters and users. However, conventional remote speech-to-text systems necessitated ensuring the availability of a network and preparing and operating devices for the system establishment and, hence, could be operated only by a limited number of organizations and groups. To overcome these limitations, we developed a system called captiOnline through which one can perform remote speech-to-text interpretation by simply accessing it through a web browser. In this paper, we explain how to implement captiOnline, evaluate its use in actual situations, and compare it with conventional systems. In addition, we show that captiOnline can realize remote speech-to-text interpretation in a simple configuration requiring less preparatory works.