Abstract: |
The goal of the dissertation is to develop a system that translates Sign Language (SL) spelled-words to text, thus aiding the hearing impaired people. A word in SL can be expressed with a single sign or a combination of signs. This fact by itself makes the sign recognition and its translation to natural language quite a complex problem. As a first phase in dealing with these issues, we focus on recognizing the SL alphabet, in which each letter is represented by a unique and discrete hand movement. The chosen approach in recognizing the sign letters is vision-based and consists in identifying the hand region in each of the frames from a video sequence of continuous alphabet signs. At the current stage of the system development, we focus on the proper hand region segmentation. |