Image pre-processor for character recognition system

An image processing apparatus for converting a gray-scale image to a binary image by thresholding, comprising:

• a memory for storing the gray-scale image and for storing instruction sequences by which the gray-scale image is processed;
• and
• a processor for executing the stored instruction sequences;
• wherein the stored instruction sequences include process steps to cause said processor to (a) obtain values for a histogram of pixel intensities of the gray-scale image, the histogram being comprised by discrete groups each containing at least plural gray levels of the gray-scale image, (b) identify the top two groups in the histogram that are separated by at least one intervening histogram group, including steps to order values for all of said plural discrete groups of the histogram so as to identify said top two groups separated by at least one intervening group (c) calculate a global threshold at an intermediate distance between said two top groups, (d) compare each pixel in the gray-scale image to the global threshold so as to binarize each pixel, and (e) output a binary image corresponding to the gray-scale image.

An image processing method for converting a gray-scale image to a binary image by thresholding, said method comprising the steps of:

• obtaining values for a histogram of pixel intensities of the gray-scale image, the histogram being comprised by plural discrete groups each containing at least plural gray levels of the gray-scale image;
• identifying the top two groups in the histogram that are separated by at least one intervening histogram group, wherein said identifying step includes ordering values for all of said plural discrete groups of the histogram so as to identify said top two groups separated by at least one intervening group;
• calculating a global threshold at an intermediate distance between said two top groups;
• comparing each pixel in the gray-scale image to the global threshold so as to binarize each pixel;
• and
• outputting a binary image corresponding to the gray-scale image.

Computer-executable process steps which convert a gray-scale image to a binary image by thresholding, said process steps including:

• an obtaining step which obtains values for a histogram of pixel intensities of the gray-scale image, the histogram being comprised by plural discrete groups each containing at least plural gray levels of the gray-scale image;
• an identifying step which identifies the top two groups in the histogram that are separated by at least one intervening histogram group, wherein said identifying step includes ordering values for all of said plural discrete groups of the histogram so as to identify said top two groups separated by at least one intervening group;
• a calculating step which calculates a global threshold at an intermediate distance between said two top groups;
• a comparing step which compares each pixel in the gray-scale image to the global threshold so as to binarize each pixel;
• and
• an outputting step which outputs a binary image which corresponds to the gray-scale image.