Distributed architectures for pen-based input and diagram recognition

Abstract

We present a system supporting pen-based input and diagram recognition that employs a personal digital assistant (PDA) as an intelligent input device for the system. Functionality is distributed between the PDA and the main computer, with the PDA performing low-level shape recognition and editing functions, and the back-end computer performing high-level recognition functions, including recognition of spatial relations between picture elements. This organization provides a number of advantages over conventional pen-based systems employing simple digitizing tablets. It provides the opportunity to use hardware specially designed for shape recognition and editing in a general diagram recognition system, it allows for improved performance through parallel processing, and it allows diagram entry to be performed remotely through use of the PDA front end in the field, with recognized shapes subsequently downloaded to the main diagram recognizer. We discuss the overall organization of the system, as well as the individual pieces and the communication between them, and describe two ongoing projects employing this architecture. © 1996 ACM.