Binarization method

A binarization method of obtaining an optimum threshold value for binarization which is used when converting multi-level image data which describes an input image into black-and-white bi-level image data, the binarization method comprising the steps of:

• a) dividing the input image into a plurality of regions;
• b) discriminating whether or not each region is a character region which constitutes a portion of a character;
• c) integrating regions which are discriminated as character regions into an integrated region when a number of regions which ate discriminated as the character regions exceeds a predetermined value:
• d) counting a first number of black picture elements by varying a threshold value from a darkest tone level to a lightest tone level;
• e) counting a second number of picture elements having a tone level darker than the lightest tone level within the integrated region;
• f) obtaining a percentage of the first number with respect to the second number for each of the varied threshold values so as to normalize a density of the input image;
• and
• g) determining the optimum threshold value for binarization based on the percentage which is obtained for each of the varied threshold values;
• wherein the two counting steps d) and e), the obtaining step f), and the determining step g) are performed with respect to the integrated region.