期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2014
卷号：3
期号：10
页码：3388-3392
出版社：Shri Pannalal Research Institute of Technolgy
摘要：Hand gestures are primary means of communication among deaf and dumb humans and sign language is the most natural ,expressive and effective way of communication for dump and deaf people. This paper presents an image processing technique for finding equivalent readable text of Devnagari Sign Language. It attempts to process static images of the Devnagari sign language and then matches them to a statistical database of pre-processed images to recognize the specific set of signed letters that contains sign of vowel alphabets. We consider a constant Environment for gesture recognition. Our approach contains steps for resizing image, converting the RGB image to Binary image, removing noise from this image, segmenting the hand region, finding out its area, mean, and centroid, solidity , entropy ,correlation features from this preprocessed image. Then we create a database based on this features and classify the gesture. We use Feed-forward neural network (pattern recognition tool ) to classify and detect the signs of Devnagari Sign Language.
关键词：Devnagari Sign Language (DSL); Feed Forward ; Neural Network; patten recognition ;Gesture Recognition; ; Features Extraction
