Filtering and Reducing Dimension in the Recognition of Vietnamese Static Sign Language

Tóm tắt

Sign language is the primary language for the deaf in communicating with normal people. In this paper, Vietnamese Static Sign Language (VSL) using filtering and dimension-reducing methods combined with Neural network has been proposed. This approach begins with pre-processing steps, feature extraction and classification to recognize and to show the results of the relevant letter. In this paper, the hand’s region is segmented from the background using skin color; after that, filters are used to reduce noise/ to smooth noise before extracting the features using the down-sampling method. These features are then used to train and to test with the back-propagation Neural Network. The implementation is performed on the database that was built with some conditions. The combination of the above-mentioned algorithms based on self-built databases results in a relatively high outcome.