AiSL

Team Members:

• Elshaddai Melaku - [melakueb@mail.uc.edu]
• Ikran Warsame - [warsamih@mail.uc.edu]
• Fareena Khan - [khanfe@mail.uc.edu]

Advisor: Dr. Jillian Aurisano

---

Table of Contents

1. Project Description

   • Project Description

2. User Stories and Design Diagrams

   • User Stories
   • Design Diagrams

3. Project Tasks and Timeline

   • Task List
   • Milestones, Timeline, and Effort Matrix

4. ABET Concerns Essay

   • ABET Concerns

5. Presentation

   • Fall Presentation
   • Poster
   • Final Presentation

6. Self-Assessment Essays

   • Elshaddai: Contract Final Self-Assessment
   • Ikran: Contract
   • Fareena: Contract | Final Self-Assessment

7. Professional Biographies

   • Elshaddai Bio
   • Ikran Bio
   • Fareena Bio

8. Summary of Hours and Justification

   • Summary of Hours

9. Budget

   • Summary of Expenses

10. Appendix

    • Appendix Folder
    • Initial Brainstorm
    • Idea 1 Brainstorm
    • Resources
    • Considerations

---

Project Abstract

AiSL is an AI-powered American Sign Language learning tool that converts English sentences into ASL grammar and plays them back through a 3D signing avatar built in Unity. The system prioritizes linguistic accuracy over direct word-for-word translation, using a retrieval-augmented LLM pipeline to generate structured ASL gloss output. Learners can observe how English sentence structure differs from ASL, with playback controls for speed and looping. The project was developed with input from the Deaf community and is designed as an accessible, transparent educational prototype rather than a replacement for native ASL instruction.

---

Credits and Data Sources

• 3D Avatar: Microsoft Rocketbox
• 3D Animation Clips: Studio Galt Sign Language Mocap Archive
• ASL Sign Videos: WLASL Processed Dataset (Kaggle)