Speech and Language Technology for Disabled Persons
Three different approaches for creating AutoCAD drawings with voice input have been examined: isolated-word speech recognition, connected-speech recognition, and word prediction (followed by translation into drawings). Commercially available isolated-word recognizers are limited by a small vocabulary size and by training requirements. Connected-speech recognizers permit faster and more natural interaction. Word prediction can be used to specify textual rules which facilitates modification of drawings.
Bibliographic reference. Bickley, Corine / Hunnicutt, Sheri / Lamel, Lori (1993): "Alternative strategies for creating autoCAD drawings", In SLTDP-93, 103-106.