An improved framework for layering linguistic processes in sign language generation: Why there should never be a “brows” tier

Abstract

Creating legible animations of sign language that satisfy the needs of native users requires drawing from the fields of computational linguistics, computer animation and user experience research. This paper explores the problem of layering the many linguistic processes that contribute to a sign language utterance in avatar animation. A new framework is presented that satisfies requirements from each of these fields, and yields a flexible architecture for sign language avatars capable of leveraging a wide range of animation techniques to generate rich multi-layered animations of sign.