Input and output modalities used in a sign-language-enabled information kiosk

This paper presents description and evaluation of input and output modalities used in a sign-language-enabled information kiosk. The kiosk was developed for experiments on interaction between computers and deaf users. The input modalities are au-tomatic computer-vision-based sign language recognition, au-tomatic speech recognition (ASR) and a touchscreen. The out-put modalities are presented on a screen displaying 3D signing avatar, and on a touchscreen showing special graphical user in-terface for the Deaf. The kiosk was tested on a dialogue provid-ing information about train connections, but the scenario can be easily changed to e.g. SL tutoring tool, SL dictionary or SL game. This scenario expects that both deaf and hearing people can use the kiosk. This is why both automatic speech recogni-tion and automatic sign language recognition are used as input modalities, and signing avatar and written text as output modal-ities (in several languages). The human-computer interaction is controlled by a computer-driven dialogue system. 1