Sign language interpretation using linear discriminant analysis and local binary patterns
Mahmood Jasim, Mohammed Hasanuzzaman · 2014
This paper presents a computer vision-based hand sign gesture recognition system for sign language interpretation. Haar-like feature-based cascaded classifier is used for hand area detection. Hand gestures portraying sign language are recognized using Linear Discriminant Analysis and Local Binary Pattern based feature extractors separately. The sign gestures are classified using Nearest Neighbor algorithm. For testing the system the Chinese and Bangladeshi Numeral Gesture datasets are prepared containing sign gestures describing the numerals of 0 to 9 for the respective languages. The mean accuracy of LDA based sign language interpretation on the Chinese numeral gesture dataset is 92.417% and on the Bangladeshi numeral gesture dataset is 88.55%. The mean accuracy of LBP based sign language interpretation on the Chinese numeral gesture dataset is 87.13% and on the Bangladeshi numeral gesture dataset is 85.10%.