Framework of ASL silhouette gesture recognition system

Abstract

The Sign language is an approach to impart for hard of hearing individuals, which rigid shapes are utilized as a sound patterns. In this paper, we exhibit a system intended for identifying letters and the numeral’s in the American Sign Language in light of saliency identification of the images. In the wake of identifying saliency, the images were handled by Independent Component Analysis (ICA), with a specific end objective is to decrease measurements and expand the class internal similitude and diminish class external resemblance. At final resultant vectors will be taught and classified through support vector machine (SVM).The utilization of this framework in the communication of the deaf people and in addition toward connecting with the computer, this is because of the utilization of standard letters in the sign language. The acknowledgment rate of the framework was 99.92% utilizing 4-fold cross validation method in which training conditions lying on the average. The consequences of the proposed system speak to high exactness and legitimate execution of this system compared among the others.