Recognising sign using fractional energy of cosine haar hybrid wavelet transform of american sign images
Archana G. Funde, Sudeep D. Thepade · 2016
The communication is the ability that makes the basic quality that differentiate human beings. Hand gesture, facial expression and finally the oral medium are the communication medium. Sign language is most important medium of communication for deaf and dumb people, it also helps the military area, online forums and security purposes. In this paper sing language Recognition is done using various transform and Cosine-Haar Hybrid Wavelet Transform generated using Assorted proportion of constituent Transform and various similarity measured like MSE, Abs, Canberra, Manhattan, Sorensen, square-chord, Taxial Distance etc. Sign language recognition and edge detection is major challenge for object detection using gradient operation and various transform. The paper deliberate about slope magnitude method for detecting edge and multiple transform for energy composition and also extract the feature by using column transformed image, Going ahead with the help of these edge image and various transform calculate the GAR. (Genuine acceptance ratio). Cosine Haar hybrid gives the better performance other than individual transforms.