American Sign Language Detection and Text to Speech Conversion -Towards Inclusivity
Hitha B Mendon, Savi Sanjiv, Us Priyanka, Shetty Samarth Gunapa, Manjula M. Ramannavar · 2025
Speech impairment may lead to social exclusion where its victims are kept isolated with feelings which negatively affect their morale as is demonstrated on these disabled populations. The system therefore should intend developing an automated solution translating an American Sign Language (ASL) gesture into either the text or speech of people in real time in this aspect: aiming for social inclusion. The system utilizes Mediapipe for hand gesture detection, OpenCV for image preprocessing, and Pyttsx3 for text-to-speech conversion. Mediapipe extracts key landmarks from ASL gestures captured through webcam, while OpenCV ensures accurate gesture recognition. A machine learning model, trained on a comprehensive ASL dataset using the Random Forest algorithm, maps gestures to text, which is subsequently converted into speech using Pyttsx3. The engineered system is very robust, with a remarkable model accuracy of 99.8% in recognizing ASL gestures and translating them into understandable speech. This ensures high reliability and ease of use in practical applications. This work demonstrates how technology can help people with speech disabilities overcome communication barriers and be included in the mainstream. It’s a step toward an open and fair communication platform where social and professional integration could take place better.