Japanese JSL Translation and Searching Display Conditions for Expressing Easy-to-Understand Sign Animation

This paper described a bi-directional translating system between Japanese and Japanese Sign Language and two experiments conducted to clarify necessary conditions for displaying easy-to-read animation of a person model speaking sign language synthesized by the system. In Experiment 1 hearing-impaired and hearing subjects judged identity of a pair of animation. Resolution and frame rate had an main effect against correct answer rate. While the hear ing had tendency to watch the whole body of the model, the hearing-impaired focused their attention chiefly on handshape and movement. In Experiment 2 the hearing-impaired read signs on animation. For correct reading above 90 % sign animation needed to be displayed with more than 10 x 8 cm, 113 x 90 pixels and 8 f/sec. This condition was much relaxed compared with that in Experiment 1.