Statistical machine translation for Indonesian-German sign language

Abstract

In the world of sign language, development has yielded significant results in recent years. Much of the research in the field of sign language has been done. Mostly, the research has designed for signerindependent schema that contains sign language and video corpus of signers. Although there are many studies in the field of sign language, but rarely there is research to translate from sign language to another sign language. This paper, contributes to the research about translating words from Indonesian-German Sign Language. For the translation process a Markov model and the parsing tree method to translate the words are used. Two different methods namely a binary search algorithm and a binary search algorithm with Markov model are used for the process of translating a word which derived from the input. A database that contains the list of words used in daily conversations is prepared. A binary search algorithm and Markov models given a better result when translated single words when are compared with a binary search algorithm. However, binary search algorithm give a good result two or more words are translated.