Sign Language Decoder for the Hearing Impaired

Abstract

Sign language, the sole medium of contact for dumb and deaf people. These visually disabled people communicate their feelings and ideas to other persons with the aid of sign language. As it is difficult for the normal human to understand these languages, these mentally disabled individuals always have to hold the interpreter along for connecting with the world. Recognition of sign’s has since become an analytical assignment. Since sign language contains a variety motions and hand gestures, the consistency of sign’s relies on the exact recognition of hand gestures. This paper presents a novel approach for sign language recognition. The primary aim is correctly classify and identify the hand gesture. The following work consists of segmentation models concerning skin color for areas of hand and detection from the background, HOG method is used, extracts the shape information features and finally Random Forrest machine learning classifier is utilized with effective results. The accuracy of our new models was It was revealed to be considerably higher than existing one.