GSLC: Creation and annotation of a greek sign language corpus for HCI

Abstract

In the framework of a research target that aims at integration of sign language technologies to human-computer interaction applications, creation and annotation of the Greek Sign Language Corpus (GSLC) involve, on the one hand, data and analysis of the phonological structure of morphemes of Greek Sign Language (GSL) and, on the other hand, collection of sentence level language samples and assignment of their respective annotations. GSLC also entails free sign narrations fully annotated at least for sentence segmentation. Simple and complex sign morpheme formation is directly relevant to development of sign recognition prototypes. In this sense, a sign language corpus intended to support sign recognition by exploitation of a language model has to entail sufficient data from simple- to complex- morpheme level. Sentence level annotation, except for sentence boundaries, focuses on phrase boundary marking and grammar information often conveyed by multi-layer markers, as is the case of e.g. topicalisation, nominal phrase formation, temporal indicators, question formation and sentential negation in GSL. © Springer-Verlag Berlin Heidelberg 2007.