Hand Gesture Recognition

Abstract

Over nine billion people worldwide are blind and deaf, and 2.78 percent of Indians are physically handicapped (deaf and dumb). Since that communication is a vital part of human life, this causes them several issues. The deaf-dumb have their own visual language known as sign language. Between normal and deaf or dumb individuals, sign language is a natural means of communication. Normal individuals occasionally struggle to read the signs correctly and comprehend what they are trying to express. In this paper we propose a software system using CNN to capture transform hand gestures into audio and text so that both regular people and handicapped people may understand the recognized gesture. The proposed software system takes pictures of hand gestures with a camera and pre-processes them to extract characteristics. The CNN model is then used to classify the hand movements using these characteristics. The audio output is obtained that can be heard through a speaker. This makes the project a complete solution for deaf and dumb people because commands can be wirelessly heard and displayed.