Abstract:
Human beings have natural ability to see, listen and interact with its surrounding. Unfortunately, there are some people in the society which do not have ability to use their senses to the best extent possible. As communication is a fundamental aspect of human life, such people depend on the other means of communication like sign language (a nonverbal form of intercourse). To bridge this communication gap, it is important to have an advance sign-language detection and gesture recognition system for people in the community when they try to engage in interaction with normal public that do not understand sign-language. It is a basic necessity for every human being to share their feelings and emotions without facing any restrictions. Therefore, an effort has been made to develop a smart glove using different hardware modules and software tools for real-time gesture recognition. It is known that every person hand has unique shape and size, we aimed to design a device that could provide reliable translations regardless of those differences. The objective is to create an easy wearable device which help speech impaired people to interact with external environment as normal public do and removes communication barrier among them. The system will be capable of recognizing hand gesture and its translation into speech (audible sound) and visual text. To make this communication possible several components are integrated on the glove like sensors, microcontroller and other hardware components for real time transmission of hand gestures to Machine Learning models in python script for its translation. For system efficiency, a huge amount of data is collected and trained over different machine learning models which further experimentally examined and compared in order to select the best model among all based on the accuracy rates.
