METHOD FOR COMPRESSING SCANNED COLORED AND GRAY-SCALE DOCUMENTS

摘要

A new method for compressing scanned documents. Starting with the division of the document into three planes - foreground image, background image and mask image - the basic course of the compression method is as follows: A locally variable threshold value image is generated from the original document using an adaptive threshold method. The original image is then quantized using this threshold value image generating a bitonal quantized image. A text detection (segmenting) method uses the bitonal quantized image and the original image to divide the document into foreground and background regions, classifying text and graphics as foreground regions and images and text background as background regions. A mask identifying the foreground and background regions is generated by the text detection method. A reduced resolution foreground image, describing the colore of the foreground regions, is produced from the original image and the mask image. A reduced resolution background image, describing the color of the background regions, is produced form the original image and the mask image. The mask image, the background image and the foreground image are finally coded independently using suitable image colors.