A dipole receptive fields-based document enhancement system

Summary form only given, as follows. Since there may be defects and noises in some documents generated from fax machines or photocopiers, the authors have proposed a mechanism to improve the quality of prints for any set of fonts. A neural network model has been developed using the concept of dipole receptive fields to revise the defects of a given pattern. This algorithm consists of a preprocessing stage, sensitivity sensor, and connecting machine, and has been demonstrated to improve the quality of a printed document.<<ETX>>