Animations from Article in ACM Transactions on Accessible Computing (TACCESS)
Matt Huenerfauth, Liming Zhou, Erdan Gu and Jan Allbeck. 2008. Evaluation of American Sign Language Generation by Native ASL Signers. ACM Transactions on Accessible Computing.
This sample video is from our journal article presented in TACCESS in 2008. The video shows an animation produced using our software for generating ASL classifier predicate expressions. In this example, animations of a 3D scene being described are shown. The videos include a "Signed English" version of the description and an ASL classifier predicate version of the description. The "Signed English" version is shown for comparison purposes because it represented the state-of-the-art in automatic English-to-Sign translation/transliteration at the time. Our classifier predicate ASL animations are meant to demonstrate how our generation architecture is sufficiently expressive for conveying complex ASL spatial phenomena. Native ASL signers evaluated these animations in an experiment, and the ASL classifier predicate animations from our software were more effective at conveying information content to these users.