Document-Level positioning of a pen tip by retrieval of image fragments

This paper presents a new method of positioning a pen tip on a document for pen-based computing. Current technologies such as the Anoto pens rely on special mechanisms for positioning, which result in limiting their applicability.Our proposal is to ease this problem - a camera pen that allows us to write on ordinary paper for positioning. A document image retrieval method calledLLAH is tuned to position the pen tip efficiently on the coordinates of a document only by capturing its tiny image fragment. In this paper we report results of preliminary experiments about recovering handwriting of simple figures and letters on documents using a prototype camera pen.