Method for background suppression in an image data processing system

Claim 1

A method of suppressing background data from gray level document image data which has been captured from a document as a plurality of digitized picture element (pixel) gray level values ranging from a darkest gray level value to a whitest gray level value, the image pixels arranged in a plurality of contiguous scan lines extending from a bottom to a top edge of the document image, the method comprising the steps of: (a) selecting one of the plurality of pixels for examination; (b) defining a two-dimensional window of pixels centered about the selected pixel; (c) determining a gray level signal threshold and a gray level white threshold as functions of a gray level background estimate of at least a preselected pixel in the window; (d) determining counts of (1) all window pixels having gray levels in a range of gray levels extending between and including the white threshold and the whitest gray level and classifying such pixels as white pixels, (2) all window pixels having gray levels in a range of gray levels extending between and including the signal threshold and the darkest gray level and classifying such pixels as signal pixels, and (3) all window pixels having gray levels between the white threshold and the signal threshold and classifying such pixels as background pixels; (e) determining (1) a signal average gray level by summing gray levels of all signal pixels in the window and dividing by the count of the signal pixels, (2) a white average gray level by summing gray levels of all white pixels in the window and dividing by the count of white pixels, and (3) a background average gray level by summing gray levels of all background pixels in the window and dividing by the count of background pixels; (f) determining a window classification as one of all white, all signal, all background and mixed as a function of the counts determined in step (d); (g) determining a suppression threshold gray level for the selected pixel by (1) using a preselected function of the background average whenever the window classification is determined to be all background, (2) using a preselected function of the background estimate whenever the window classification is determined to be all signal, (3) using a preselected function of the window white average whenever the window classification is determined to be all white, or (4) using preselected functions of one of the white average, signal average, background average, white threshold and background estimate whenever the window classification is determined to be mixed; (h) determining a new gray level value for the selected pixel as a function of the selected pixel's current gray level value and the suppression threshold gray level value for the selected pixel; and (i) selecting another one of the plurality of pixels for examination and repeating steps (b) through (h) for every pixel in a document image.

Claim 2

A method of suppressing background data from gray level document image data which has been captured from a document as a plurality of digitized picture element (pixel) gray level values ranging from a darkest gray level value to a whitest gray level value, the image pixels arranged in a plurality of contiguous scan lines extending from a bottom to a top edge of the document image, the method comprising the steps of: (a) selecting one of the plurality of pixels for examination; (b) defining a two-dimensional window of pixels centered about the selected pixel; (c) determining a gray level signal threshold and a gray level white threshold as functions of a gray level background estimate of at least a preselected pixel in the window; (d) determining counts of (1) all window pixels having gray levels in a range of gray levels extending between and including the white threshold and the whitest gray level and classifying such pixels as white pixels, (2) all window pixels having gray levels in a range of gray levels extending between and including the signal threshold and the darkest gray level and classifying such pixels as signal pixels, and (3) all window pixels having gray levels between the white threshold and the signal threshold and classifying such pixels as background pixels; (e) determining (1) a signal average gray level by summing gray levels of all signal pixels in the window and dividing by the count of the signal pixels, (2) a white average gray level by summing gray levels of all white pixels in the window and dividing by the count of white pixels, and (3) a background average gray level by summing gray levels of all background pixels in the window and dividing by the count of background pixels; (f) determining a window classification as one of all white, all signal, all background and mixed as a function of the counts determined in step (d); (g) determining a suppression threshold gray level for the selected pixel by (1) using a preselected function of the background average whenever the window classification is determined to be all background, (2) using a preselected function of the background estimate whenever the window classification is determined to be all signal, (3) using a preselected function of the window white average whenever the window classification is determined to be all white, or (4) using preselected functions of one of the white average, signal average, background average, white threshold and background estimate whenever the window classification is determined to be mixed; (h) determining a new gray level value for the selected pixel as a function of the selected pixel's current gray level value and the suppression threshold gray level value for the selected pixel; and (i) selecting another one of the plurality of pixels for examination and repeating steps (b) through (h) for every pixel in a document image.