A Neural Network based system for Persian sign language recognition
Azadeh Kiani Sarkaleh, Fereshteh Poorahangaryan, Bahman Zanj, Ali Karami · 2009
This paper presents a static gesture recognition system for recognizing some selected words of Persian sign language (PSL). The required images for the selected words are obtained using a digital camera. The color images are first resized, and then converted to grayscale images. Then, the discrete wavelet transform (DWT) is applied on the selected images and some features are extracted. Finally, a multi layered Perceptron (MLP) Neural Network (NN) is trained to classify the selected images. Our recognition system does not use any gloves or visual marking systems. The system was implemented and tested using a data set of 240 samples of Persian sign images; 30 images for each sign. The experiments show that the proposed system is able to classify the selected PSL signs with a classification accuracy of 98.75% when the network is trained using MATLAB NN Toolbox.