The Method of Generalized Grammar Structures for Text to Gestures Computer-Aided Translation
- PDF / 259,190 Bytes
- 8 Pages / 594 x 792 pts Page_size
- 106 Downloads / 161 Views
THE METHOD OF GENERALIZED GRAMMAR STRUCTURES FOR TEXT TO GESTURES COMPUTER-AIDED TRANSLATION Iu. V. Krak,a O. V. Barmak,b† and S. O. Romanyshyn b‡
UDC 004.9
Abstract. An approach to solving the problem of translation to sign languages is described. Generalized grammatical structures are developed based on sentence corpus used by hearing impaired people in regular communication. Practical implementation of infological models of inflected and sign languages is described. A translation algorithm is developed. The experimental confirmation of the approach is described based on the Ukrainian language. Keywords: computer-aided translation, text, sign language, infological model. INTRODUCTION We will consider sign language (SL), a natural language passing information by means of hands and fingers gestures, facial expressions, and posture of the person. This language is used as a constituent for dialog and serves as the main method for deaf people communication [1, 2]. Sign languages are not visual interpretation of normal languages, instead they possess the grammar and can be used for discussion of various subjects: from simple and specific to esoteric or abstract. The lexical characteristics of SL are still insufficiently studied, especially idiomatics, phraseology, and morphology. Automatic attributing of communicational SL to all forms of verbal and written speech is incorrect. Communicational SL used by deaf people cannot be studied utilizing traditional lingvistical models. The most appropriate way to study of SL morphology is by using the principle: from value to the form. In view of it the problems arising at creation of computer-aided translation systems of an arbitrary text in SL consist in determination of complementary pairs of structures that convey the meaning of information. Computer-aided translation [3] differs from machine translation [4, 5] in that it assumes the following forms of interaction with the user: partially automatic translation (usage by the person-translator of computer dictionaries); systems with division of labor in which the computer is trained to translate only rigidly structured phrases (without the subsequent corrections), and a phrase with different structure is translated by the person. Let us consider syntactic features of SL on an example of three typical sentence structures: subject–object–verb; subject–verb–object; verb–subject–object. The subject and the predicate in such sentences are connected predicatively. We will call sentences with one predicative connection simple sentences. Note that the word order in sentences with one predicative connection in the majority of spoken languages of the world is described by one of three specified typical structures [6]. In SL simple sentences serve as the main mean of communication and are divided into narrative, interrogative and imperative [7]. For implementation of a computer-aided translation of the initial text information in SL, assume that the following pairs are possible: a sentence in a source language – analog in SL in the for
Data Loading...
