Russian sign language machine interpreter system based on the analyses of syntax and semantic construction

Abstract

In this paper the existing systems of translation into sign language are analyzed. There are ViSiCAST project, Zardoz system, ASL Workbench, TEAM system. Their strengths and weaknesses are identified. A new translation way based on a comparison of syntactic constructions is described. Sentences in sounding language often describe complex situations, including some simple situations (for example, a few actions). The unit, which is processed about, is a proposal for a full participle. As a result of comparison of syntactic constructions is obtained a set of simple sentences in which all participles replace verb. A corresponding library to determine the syntax construction is developed. A new method for constructing a semantic unit of the computer system of sign language is proposed. The result of semantic analysis system is a list of correspondence “word-gesture”. Among the many alternatives of words based on semantic analysis algorithm, every word is assigned a unique lexical meaning. For simple sentences algorithms of semantic analysis are designed and implemented. The most priority directions of modification module semantic analysis may include the following: broadening the base of gesture, the implementation of parsing complex sentences, taking into account specifics of the Russian-language.