Pattern matching for automatic sign language translation system using LabVIEW

This paper presents an automatic sign language translator, which is able to translate Malaysian sign language using pattern-matching algorithm. The sign language translator is a vision-based system where the image of the sign is captured by a camera, processed and translated into English by the computer. This sign language translator is able to recognize alphabets (A-Z), numbers (0-9), finger spelling, words (13 words) and sentences. Alphabets, numbers and fingers are categorized under static signs while words and sentences are known as dynamic signs where the signs involve motion. Static signs are recognised by matching positions of each fingertip with the database while the recognition of dynamic signs is performed by comparing the trajectory of the motion. The accuracy for static sign is 97.79% while the accuracy for dynamic sign is 80.38%.