Abstract:
A pre processing method for table form document image in document image processing is discussed in this paper. Because the box frame in the table form document touches or overlaps the characters in the box frame, which will affect the distinguishability of an OCR system, the performance of the total system drops. In order to solve this problem, a basic and general model of line style noise removal and its detailed algorithm are proposed. Based on this model, some extended methods are also proposed for special applications. Some experiments are also given to show the effectiveness of these methods.