A Machine Learning Algorithm to Assist People with Speaking Disabilities for Sign Language Detection of People Speaking Uzbek Language

This paper describes a current work that aims to fulfill the communication needs of those who are speech impaired, with a particular emphasis on Uzbek sign language users. Using machine learning techniques, the work attempts to recognize hand motions in real time and then uses Python, OpenCV, MediaPipe, and scikit-learn to translate them into spoken Uzbek. The Random Forest classifier achieves respectable accuracy rates in spite of the difficulties that come with curating datasets. The current version demonstrates the capacity to record hand gestures in real time, accurately read sign language, and produce spoken words. This research work offers a glimpse of the work in progress, with future plans to extend the functionality of the program to include gesture detection in real time. The knowledge gained from this will likely advance the development of assistive devices by providing optimistic prospect about the potential to enable efficient sign language communication.