Development and testing of the algorithm of semantic analysis of speech (text) for translation into Russian Sign Language

Abstract
The paper makes analysis of some current computer sign language interpretation systems. Their advantages and disadvantages are revealed. The main disadvantage is the lack of an original text semantic analysis unit capable to solve the problem of homonyms. A general scheme of translation system from sounding Russian into Russian Sign language, including semantic analysis unit, is presented. The block of source code analysis, taking into account the semantic component of the Russian language, is proposed. The work of the semantic module is based on Tuzov’s dictionary. A semantic analysis algorithm is described. The text analysis isn’t completed until each word has only one semantic description thus solving the problem of ambiguity. The major features of the new semantic analysis module include the following: expanding the base of gestures, implementing parsing complex sentences, adding the account in the algorithm analyzer of classifier predicates of Russian Sign Language. Testing of the algorithm is made.

Keywords: Russian Sign language, computer sign language interpretation, semantic analysis, morphological analysis, syntactic analysis, homonyms, syntax constructions, alternatives.