SLRMPCMC: Sign Language Recognition using Mediapipe and Cross-model Comparison

Sign language is key for the deaf or hearing-impaired, but traditional reliance on interpreters presents limitations in both availability and precision. Historical methods for sign translation faced unique challenges from the visual-spatial aspects of sign languages, often failing to accurately recognize gestures or adapt to varying contexts and necessitating instant translation. This caused constraints in usability and hindered the application on diverse platforms such as mobile devices due to high computational requirements. This research introduces a sophisticated framework for sign language recognition, integrating the Mediapipe Landmark system with ab combination of machine learning models—Long Short-Term Memory networks, Convolutional Neural Networks, and Random Forest— to enhance the precision of detecting hand signs and movements in real-time and across different settings. The study outlines the processes for data gathering, preprocessing, and classifying gestures and static signs, advancing communication technologies for those with hearing impairments.