A Dynamic Finger Character Recognition Method Using Landmark Behavior Rule Base

Abstract

This paper proposes a dynamic finger character recognition method using a landmark behavior rule base. By applying this rule base, not only static finger characters but also dynamic finger characters can be recognized, and all Japanese finger characters can be recognized. Japanese finger characters are finger character representations with hand finger shapes of the Japanese call letters in the Japanese language. The two main types of finger character forms are. One is the static finger character, which is represented by the shape of the hand without any motion, and the other is the dynamic finger character, which is represented by the shape of the hand and motion. Our method realizes recognition of both the static and dynamic finger characters, and continuous recognition. To achieve more flexible finger character recognition, it is required to recognize not only specific characters, but also all characters that can make up a word. The realization of automatic translation of sign language or finger characters will facilitate communication and social participation between both sign language and oral language speakers.