Automatically Generating Sketch Interfaces from Shape Descriptions

As pen-based input devices have become more common, sketch recognition systems are being developed for many hand-drawn diagrammatic domains such as mechanical engineering, GUI design, course of action diagrams, and many others. These sketch interfaces 1) allow for more natural interaction than a traditional mouse and palette tool by allowing users to hand-sketch the diagram, 2) can automatically connect to a CAD system preventing the designer from having to enter the same information twice, 3) can offer realtime design advice from CAD systems, 4) allow more powerful editing since the shape is recognized as a whole, 5) provide diagram beautification to remove mess and clutter, 6) use display as a trigger to inform the user that the shapes have been correctly recognized. However, sketch recognition systems can be quite time consuming to build if they are to handle the intricacies of each domain. Also we would prefer that the builder of a sketch recognition system be an expert in the domain rather than an expert in sketch recognition at a signal level. Rather than build each recognition system separately, our group has been working on a multi-domain recognition system that can be customized for each domain.