Elly Savatia on AI and Motion Capture for Kenyan Sign Language

Bridging Communication Gaps for the Deaf Community

• 💡 Ellie Savatia, a social entrepreneur from Nairobi, founded Signvrse to address communication challenges faced by the deaf community.
• 🎯 The company specializes in research at the intersection of sign language and technology to bridge gaps between deaf and hearing individuals.

Terp360: AI-Powered Sign Language Translation

• 🚀 Terp360 is Signvrse's software platform that interprets spoken voice and text into Kenyan Sign Language using AI, motion capture, and 3D avatars.
• 🗣️ This system aims to provide more accurate and human-like sign language articulation, including facial expressions, moving beyond basic video conferencing tools.
• ⚠️ A key challenge was the lack of Kenyan Sign Language data sets, as most public data focused on American or British sign languages.

Overcoming Data Set Challenges with Motion Capture

• 🇰🇪 Signvrse secured a grant from the government of Kenya to invest in motion capture systems, solving the data set problem.
• 🧑‍💻 By working with deaf individuals and using motion capture suits, they collect detailed body data for signing specific words and phrases.
• 📊 This collected data is then processed to enable the platform to generate corresponding avatar motions for input text and speech.

Future Growth and Empowerment

• 🏆 Signvrse was recognized among 20 organizations selected for the Google AI Accelerator, which will help improve their data collection efficiency.
• 📈 The company has a target of creating a million signs to expand their translation capabilities.
• 🤝 Savatia emphasizes the importance of involving persons with disabilities in developing solutions, adhering to the mantra "nothing for us without us."