Abstract
This paper proposes an interactive Japanese sign language interface. The proposed interface is useful to person who has difficulty in hearing. The interface has three functions explained; 1) Japanese sign language motin recognition, 2) Japanese sign language motin generation with animation, 3) Japanese sign language motion learning. Function 1) is realized by a fuzzy label matching method by means of associative inference, for unspecified human motion. Function 2) is realized by an associative inference's interactive inference feature. Function 3) is realized by an activation bidirectional propagation learning method. We show proposed interface usefulness from experiment results. We can apply the proposed interface to nonverbal interfaces.
