A Real-Time System for Classification of Pakistan Sign Language using Machine Learning

Sign language is used by the deaf to interact with hearing individuals in the community. Despite the fact that sign language is often used, the general population is unaware of it, and only hearing-impaired people are familiar with its usage. In this article, we have created a system for real-time sign language recognition in order to communicate with folks who are deaf or hard of hearing by using Pakistan Sign Language (PSL) as its sign language. Ten types or indications of PSL are differentiated in this study. Flex sensors in the glove can detect the position of certain fingers. The MPU-6050, a micro-electro-mechanical device is used to capture the unique characteristics of the hand's movement in a 3D space. In this setup, a Raspberry Pi 3B serves as the microcontroller. MPU-6050 and Flex sensors performance are integrated with additional elements including mean, energy, variance, and correlation for classification. In our investigation, we have collected information from over 3,000 movement feature samples from different people. To increase the precision of the proposed methodology, we have incorporated several machine learning classification algorithms.