Designing a User-Centric American Sign Language Translation System: Integrating Machine Learning with Design Thinking

There is a growing interest in American Sign Language (ASL) in Thailand, where technologies for real-time ASL detection, translation, and learning remain underexplored. This study presents “Manu,” a novel system leveraging machine learning and design thinking to create a user-centric ASL translation and learning platform. Manu combines machine learning algorithms, cloud technologies, and a React-based frontend for real-time translation, interactive lessons, and progress tracking. Usability studies with ASL experts and general users validated its effectiveness, showing a mean System Usability Scale (SUS) score of 74.69 and an average translation accuracy of 81.24%. Results demonstrate good usability, high user satisfaction, and strong intentions to use the system. Ongoing enhancements aim to optimize translation speed, accuracy, user interface, and mobile accessibility, showcasing the application of design thinking in sign language education. This study highlights Manu's potential to enhance ASL education and bridge communication gaps.