Dianoema: Optical Analysis and Recognition for modeling sign language

Start Date: 13/06/2006
End Date: 12/12/2007
Funding: Information Society - Measure 3.3
Project Leader: Fotinea Stavroula-Evita

The objective of project DIANOEMA is the development of a system of recognition of the Greek Sign Language through visual video analysis. For the needs of the project, innovative image analysis and computational vision algorithms will be developed for effective analysis of video sequences for tracking and observing gestures, and, at the same time, a corpus of Greek Sign Language (GSL) and an annotation and modelling of its subgroups will be collected. During the first phase of the project, a gesture recognition system of the sign language will be developed and also linguistic modeling of these gestures in order to test their application as interfaces in a robot machine-pilot system.

In the next step a system for automatic recognition of GSL gesture sequences will be developed with specific grammatical significance using automatic systems in computational vision which are pre-trained in the annotated corpus video-text in order to create the original conditions for recognizing natural sign language.