Event Title

American Sign Language Translation Glove

Presenter Information

Jeffrey Herring

College(s)

College of Sciences

Submission Type

Oral Presentation

Description

American Sign Language (ASL) is a widely used method of communication for the hearing impaired across North America and Canada. While ASL is a robust and elegant way of communication, there are many instances where the deaf and hearing may struggle to communicate with each other. Additionally, there exists the need for an educational tool to learn ASL. The goal of this project is to design an electronically aided translation device in conjunction with machine learning methods in hopes of aiding those wanting to communicate using ASL. The design utilizes a variety of low-cost sensors and modules integrated onto a glove and built for use on the Arduino platform. A signed letter, word or phrase is then output via a graphical user interface. Thus far, data was obtained from five individuals who repeatedly signed each letter of the ASL alphabet using our translation glove. This data was then pre-processed and sorted into a large database which was used to train a Random Forest machine learning classifier. The initial results show that this machine learning model has an accuracy score of 99.7%. As we continue to gather data and increase the vocabulary of the machine learning model, it is our aim to design a product that will bridge the communication gap between the hearing and hearing-impaired communities and provide the public access to an affordable and usable device that will ultimately improve the user's quality of life.

Comments

Honorable Mention, Undergraduate Presentation

This document is currently not available here.

Share

COinS
 

American Sign Language Translation Glove

American Sign Language (ASL) is a widely used method of communication for the hearing impaired across North America and Canada. While ASL is a robust and elegant way of communication, there are many instances where the deaf and hearing may struggle to communicate with each other. Additionally, there exists the need for an educational tool to learn ASL. The goal of this project is to design an electronically aided translation device in conjunction with machine learning methods in hopes of aiding those wanting to communicate using ASL. The design utilizes a variety of low-cost sensors and modules integrated onto a glove and built for use on the Arduino platform. A signed letter, word or phrase is then output via a graphical user interface. Thus far, data was obtained from five individuals who repeatedly signed each letter of the ASL alphabet using our translation glove. This data was then pre-processed and sorted into a large database which was used to train a Random Forest machine learning classifier. The initial results show that this machine learning model has an accuracy score of 99.7%. As we continue to gather data and increase the vocabulary of the machine learning model, it is our aim to design a product that will bridge the communication gap between the hearing and hearing-impaired communities and provide the public access to an affordable and usable device that will ultimately improve the user's quality of life.