PopSign ASL V2.0: A Large Isolated Sign Language Dataset
Author(s)
Neubauer, Bill Cornelius
Advisor(s)
Editor(s)
Collections
Supplementary to:
Permanent Link
Abstract
PopSignAI is a bubble shooter-based mobile game to help hearing parents of deaf infants learn American Sign Language. Building on previous works, PopSignAI utilizes sign language recognition to provide real-time feedback on the player’s signing. PopSignAI focuses on a vocabulary drawn from the MacArthur-Bates Communicative Development Inventory (CDI), which are the first concepts that an english-speaking child would learn in America. PopSign V1.0 covers 250 out of the 563 concepts listed in the MacArthur-Bates CDI. Thus to address this gap, PopSign V2.0 covers the remaining 313 concepts not covered in PopSign V1.0. PopSign V2 introduces approximately 170,000 new videos, collecting data from 49 different native ASL signers, with each participant approximately providing 10 videos for each of the 313 new concepts.
Sponsor
Date
Extent
Resource Type
Text
Resource Subtype
Undergraduate Research Option Thesis