Vietnamese Sign Language Recognition using Cross Line Descriptors and Invariant Moments

Trong-Nguyen Nguyen, Huu-Hung Huynh, Jean Meunier


Sign language is the primary language used by the deaf community in order to convey information through gestures instead of words. Many researches about this language have been done, and they play an important role in life. Currently, most of the hard-of-hearings in Vietnam are facing many difficulties in community integration because of their limited ability. So we propose an approach which can recognize Vietnamese sign language, based on digital image processing combined with a machine learning method. After pre-processing, we use a combination of cross lines descriptors and invariant moments to extract the features, and then the gesture is recognized using support vector machines. This is one of the few studies on sign language applied to Vietnamese alphabet (the number of words is higher and more complex than international standards with several accented letters). The proposed approach has been tested with high accuracy and is promising.


Keywords: gesture; sign; skin color; cross line; invariant moment; kernel function; one-against-all; one-against-one.

Full Text:




  • There are currently no refbacks.

Copyright (c) 2016 International Journal of Advanced Research in Computer Science