Sign Language to Text Converter Using Hand Gestures

Abstract

Sign Language is been used by hearing impaired people who are unable to interact through speech with normal people. According to the WHO study currently around more than 1.5 billion people live with hearing loss all over the world. Without sign language it becomes hard to interact with the hearing-impaired person as common people are unaware of sign language and gestures. Hence communication becomes hard. There is a need for a intermediate to translate what they want to express. Our work aims to improve the communication with the hearing-impaired person. This paper does American sign language recognition using google’s Mediapipe Hands API to detect landmarks on the hands to detect gestures and uses Support Vector Machine (SVM) classification algorithm for achieving high accuracy. The model works on web application. The proposed concept (sign language conversion to text, speech and image) is a breakthrough for helping the hearing-impaired community.