Collecting a Motion-Capture Corpus of American Sign Language for Data-Driven Generation Research
Temas
Detalles
American Sign Language (ASL) generation software can improve the accessibility of information and services for deaf individuals with low English literacy. The understand-ability of current ASL systems is limited; they have been constructed without the benefit of annotated ASL corpora that encode detailed human movement. We discuss how linguistic challenges in ASL generation can be addressed in a data-driven manner, and we describe our current work on collecting a motion-capture corpus. To evaluate the quality of our motion-capture configuration, calibration, and recording protocol, we conducted an evaluation study with native ASL signers.
En: SLPAT '10 Proceedings of the NAACL HLT 2010 Workshop on Speech and Language Processing for Assistive Technologies (2010) pp. 89-97