Sign Language Translation Solution Development

Achievements

Built a real-time sign language translation system supporting multiple languages, providing an inclusive communication tool through cutting-edge machine learning techniques.
Enhanced accessibility with text normalization, speech-to-text features, and support for internationalization in over 107 languages, making the platform globally adaptable.

Multi-Modal Translation: Developed sign-to-text and text-to-sign translation capabilities using TypeScript, integrating machine learning models with a focus on accuracy and latency reduction.
Interactive 3D Avatar Visualization: Leveraged HumanGAN technology and 3D avatar animations for realistic and engaging sign language representations, enhancing accessibility for users with hearing impairments.
Cross-Platform Usability: Designed a seamless user experience across both desktop and mobile platforms, with a responsive interface that supports speech-to-text, video output sharing, and real-time translation updates.
Global Language Support: Implemented a robust internationalization layer with support for 107 languages, ensuring the platform’s versatility and effectiveness for users worldwide.
Skeleton Viewer for Motion Analysis: Incorporated a Skeleton Viewer for precise analysis and visualization of hand and body movements, improving translation accuracy and user feedback mechanisms.

This is my Graduation Thesis. You can see the full report in Vietnamese here.