Automated finger spelling by highly realistic 3D animation |
| |
| Authors: | Nicoletta Adamo-Villani Gerardo Beni |
| |
| Institution: | Department of Computer Graphics Technology, Purdue University; Department of Electrical Engineering, University of California |
| |
| Abstract: | We present the design of a new 3D animation tool for self‐teaching (signing and reading) finger spelling—the first basic component in learning any sign language. We have designed a highly realistic hand with natural animation of the finger motions. Smoothness of motion (in real time) is achieved via programmable blending of animation segments. The hand is utilised by a programme that automatically converts text to finger spelling with controllable playback speed and playback views. The programme can be operated in two modes for two basic applications: learning to read (finger spelling) and learning to sign. For both modes, multi‐sided views are provided. In addition, for the signing mode, a mirror view is provided for the common technique of using a mirror for feedback in practising finger spelling. An additional view revealing the joint structure of the hand allows the signer to practise the subjective view without having to guess the position of the fingers. This is the first example of highly realistic 3D animation that can be used practically to teach a basic aspect of sign language. The method is applicable more generally to sign language, and this finger‐spelling application should be regarded as a first step toward the extension of highly realistic 3D animation to sign language in general. |
| |
| Keywords: | |
|
|
