A vision based approach for Pakistan sign language alphabets recognition
Loading...
Date
2014
Journal Title
Journal ISSN
Volume Title
Publisher
Pensee Journal
Abstract
There is a persistent communication barrierbetween the deaf and normal community becausea normal person
has no or limited fluency with thesign language. A person with hear-impairment hasto express himself via
interpreters or text writing.This inability to communicate effectively between thetwo groups affects their
interpersonal relationships.There are about 0.24 million Pakistanis who are eitherdeaf or mute and they
communicate through PakistanSign Language(PSL). In this research work a systemfor recognizing hand
gestures for Pakistan Sign Languagealphabets in unimpeded environment is proposed. A digitalcamera is
used to acquire PSL alphabet’s images withrandom background. These images are preprocessedfor hand
detection using skin classification filter. Thesystem uses discrete wavelet transform (DWT) for
featureextraction.Artificial neural network(ANN) withbackpropagation learning algorithm is employed
torecognize the sign feature vectors. The dataset contains500 samples of Pakistan Sign Language alphabets
withvarious background environments. The experimentsshow that the classification accuracy of the
proposedsystem for the selected PSL alphabets is 86.40%.
Description
Keywords
Pakistan Sign Language (PSL), Discrete Wavelet Transform (DWT), Computer Vision, Artificial Neural Network (ANN), Skin Classification Filter, Back Propagation Learning Algorithm
Citation
Khan, N. S., Shahzada, A., Ata, S., Abid, A., Khan, Y. D., Farooq, M. S., Mushtaq, M. T& Khan, I. (2014). A Vision Based Approach for Pakistan Sign Language Alphabets Recognition. Pensee Journal, 76(3).