Inspiration
Decrease in Professional Human ASL Interpreters, Using this Website It Can Interpret From ASL to English or Vice Versa for Everyday Situations Which Professional Interpreters Can't Be There
What it does
It Takes ASL Signs From Live Camera Feed to Create a Transcript for the Alphabet and Common Phrases. It also Takes English Text and Turns Into Both Verbal and Animated Directions For Clear Communication Both Ways.
How we built it
Using Palantir as a Platform to Build the Website, Then Using Both Pilot and AI FDE to Further Enhance and Improve Functions and Data Usage to Translate.
Challenges we ran into
Hand Position Recognition - Unable or Inaccurate Recognition of Some Hand Positions Hand Tracker Recognition - Unable to Track and Obtain Data From Hand Video & Animation Demonstration - Can't Find & Embed Links, then Creating Animation Demonstrations AI Generative Error - Directions of ASL Hand Motions Different or Slightly Errored From Correct ASL Sign ASL Recognition - Unable to or Mixes Up Some ASL Signs
Accomplishments that we're proud of
Able to Recognize A Handful of ASL Signs Consistently Database of ASL Signs for Reference Hand Tracker and Data Obtained Teaching & Translating Tool
What we learned
AI - Ai can always make mistakes and isn't trustable for everything Hand Tracker - Mishaps in Recognitions Reaching the "3D" Aspect of Video Feed How to But Together a Whole Website with the Help of AI Iterations of Commands to Further Signify and Find Different ASL Signs
What's next for ASL Translator
Larger Database - Slowly Adding More and More to ASL Database Actual Machine Learning - Input Machines and Machine Learning For Better Output & Recognition
Log in or sign up for Devpost to join the conversation.