Silicon paper and a visual interface for neural networks

Abstract

An interface technology that is represented by an LCD device equipped with a transparent tablet and a pen-like pixel selector, called silicon paper, is described. The technology is advocated as a possible replacement of the current mouse-based interface. Handwritten character recognition by a neural net is an essential component of the silicon paper interface. In order to facilitate an efficient experimental environment for neural net design, a visual interface for a neural net simulator was designed and implemented. The design was used as a testbed for a new paradigm of user interface suitable for the silicon paper technology.<>