DeepASL: A Wearable System for Unobtrusive End-to-End American Sign Language Translation

 

Executive Summary

 

There is a large community of Americans (roughly one million persons in the US) with vocal communication disabilities, who would benefit from technology to improve non-verbal communication. MSU researchers have developed an unobtrusive hands-free device capable of tracking hand motion and provides sentence level translation in real time. In addition the device can be configured to translate spoken English into written text. In regards to language learning, this device may be paired with a Virtual Reality (VR) headset to provide tutoring capabilities for those wishing to learn American Sign Language (ASL).

 

Description of Technology

 

With translation services in the US totaling roughly five billion dollars in annual revenue and a relatively low concentration of companies that provide those services, there is an opportunity for a novel technology to capture a significant share of the market. This technology uses state-of-the-art software to provide real time translations from ASL to spoken English with significant accuracy. The hands-free device responsible for tracking hand motion utilizes infrared technology, eliminating dependence on lighting conditions common in many current technologies.  Translation from ASL into English is available at the sentence level, facilitating improved real-time communication. 

 

Key Benefits

  • More robust – Infrared technology removes dependence on lighting
  • Comfortable – Technology doesn’t require bulky gloves, markers or a laptop

 

Applications

  • ASL translation
  • Sign language translation
  • Sign language tutoring

 

Patent Status:

 

Patent Pending

 

Licensing Rights Available

 

Full Licensing rights available

 

Inventors: Mi Zhang

 

Tech ID: TEC2017-0097

Patent Information:

For Information, Contact:

Raymond DeVito
Technology Manager
Michigan State University
devitora@msu.edu
Keywords: