Document-Analysis Systems and Techniques
Richard G. Casey, Kwan Y. Wong · 2020
This chapter provides a survey of different document-analysis systems and techniques. Documents used to communicate information in the office can be created, distributed, and stored in paper or in electronic forms. An optical scanner for documents generally consists of a transport mechanism plus a sensing device that converts light intensity into gray-levels. Printed documents usually consist of black on a white background. Adaptive-thresholding schemes can be either global or local. A global method determines a threshold value based on an analysis of the histogram of image-intensity values. Component labeling is a well-known image-processing technique that associates adjacent black pixels to detect distinct objects in a field. Component labeling may be done either pixel-by-pixel or using run-lengths.