Business form image identification using projected profiles of graphical lines and text string lines

Claim 1

A method for identifying business forms by constructing contour polygons from linear line segments that have been derived from a windowed raster scan of an image of the business form comprising the steps of: a) detecting the binary values of the pixels within the scanning window as the window is scanned over the image; b) positioning each of the binary values into an assigned and ordered position forming a string of binary values for each window scan with each string of binary values forming an integer number; c) directing the formed integer number to a look-up-table that translates each integer number into one or more output template commands; d) joining linear line segments into at least one contour polygon in response to the one or more output template commands; e) examining each contour polygon to determine if it represents a text component or a graphic component; f) grouping all text contour polygons and graphic contour polygons; g) profiling each of the contour polygons in the text group and the graphics group; h) accumulating the profiles for each of the groups; and i) comparing the accumulated profiles of step h against a Library of stored profiles of known forms so as to determine a best match thereby identifying the unknown form.

Claim 2

A method for identifying business forms by constructing contour polygons from linear line segments that have been derived from a windowed raster scan of an image of the business form comprising the steps of: a) detecting the binary values of the pixels within the scanning window as the window is scanned over the image; b) positioning each of the binary values into an assigned and ordered position forming a string of binary values for each window scan with each string of binary values forming an integer number; c) directing the formed integer number to a look-up-table that translates each integer number into one or more output template commands; d) joining linear line segments into at least one contour polygon in response to the one or more output template commands; e) examining each contour polygon to determine if it represents a text component or a graphic component; f) grouping all text contour polygons and graphic contour polygons; g) profiling each of the contour polygons in the text group and the graphics group; h) accumulating the profiles for each of the groups; and i) comparing the accumulated profiles of step h against a Library of stored profiles of known forms so as to determine a best match thereby identifying the unknown form.

Claim 4

A method for identifying business forms by constructing contour polygons from linear line segments that have been derived from a windowed raster scan of an image of the business form comprising the steps of: a) detecting the binary values of the pixels within the scanning window as the window is scanned over the image; b) positioning each of the binary values into an assigned and ordered position forming a string of binary values for each window scan with each string of binary values forming an integer number; c) directing the formed integer number to a look-up-table that translates each integer number into one or more output template commands; d) joining linear line segments into at least one contour polygon in response to the one or more output template commands; e) profiling the height distribution of each contour polygon; f) accumulating the height distribution for all contour polygons; g) determining the value of the maximal occurrence of the height distribution; h) setting the maximal value as a threshold value; i) classifying each of the contour polygons into a text or graphic group based on its maximal value as compared to the set threshold value; j) profiling each of the contour polygons in the text group and the graphics group; k) accumulating the profiles for each of the groups; and l) comparing the accumulated profiles of step k against a Library of stored profiles of known forms so as to determine a best match thereby identifying the unknown form.

Claim 5

A method for identifying business forms by constructing contour polygons from linear line segments that have been derived from a windowed raster scan of an image of the business form comprising the steps of: a) detecting the binary values of the pixels within the scanning window as the window is scanned over the image; b) positioning each of the binary values into an assigned and ordered position forming a string of binary values for each window scan with each string of binary values forming an integer number; c) directing the formed integer number to a look-up-table that translates each integer number into one or more output template commands; d) joining linear line segments into at least one contour polygon in response to the one or more output template commands; e) profiling the height distribution of each contour polygon; f) accumulating the height distribution for all contour polygons; g) determining the value of the maximal occurrence of the height distribution; h) setting the maximal value as a threshold value; i) classifying each of the contour polygons into a text or graphic group based on its maximal value as compared to the set threshold value; j) profiling each of the contour polygons in the text group and the graphics group; k) accumulating the profiles for each of the groups; and l) comparing the accumulated profiles of step k against a Library of stored profiles of known forms so as to determine a best match thereby identifying the unknown form.

Claim 6

A method for identifying business forms by constructing contour polygons from linear line segments that have been derived from a windowed raster scan of an image comprising the steps of: a) detecting the binary values of the pixels within the scanning window as the window is scanned over the image; b) positioning each of the binary values into an assigned and ordered position forming a string of binary values for each window scan with each string of binary values forming an integer number; c) directing the formed integer number to a look-up-table that translates each integer number into one or more output template commands; d) joining linear line segments into at least one contour polygon in response to the one or more output template commands; e) accumulating, along orthogonal axes the values of all contour polygon line segments; and f) comparing the accumulated values of step e against a Library of accumulated values of known forms so as to determine a best match thereby identifying the unknown form.

Claim 7

A method for identifying business forms by constructing contour polygons from linear line segments that have been derived from a windowed raster scan of an image comprising the steps of: a) detecting the binary values of the pixels within the scanning window as the window is scanned over the image; b) positioning each of the binary values into an assigned and ordered position forming a string of binary values for each window scan with each string of binary values forming an integer number; c) directing the formed integer number to a look-up-table that translates each integer number into one or more output template commands; d) joining linear line segments into at least one contour polygon in response to the one or more output template commands; e) accumulating, along orthogonal axes the values of all contour polygon line segments; and f) comparing the accumulated values of step e against a Library of accumulated values of known forms so as to determine a best match thereby identifying the unknown form.

Claim 8

A method for identifying business forms by constructing contour polygons from linear line segments that have been derived from a windowed raster scan of an image comprising the steps of: a) detecting the binary values of the pixels within the scanning window as the window is scanned over the image; b) positioning each of the binary values into an assigned and ordered position forming a string of binary values for each window scan with each string of binary values forming an integer number; c) directing the formed integer number to a look-up-table that translates each integer number into one or more output template commands; d) joining linear line segments into at least one contour polygon in response to the one or more output template commands; e) examining each contour polygon to determine if it represents a text component or a graphic component; f) grouping all text contour polygons and graphic contour polygons; g) accumulating the bounding values of the contour polygons of each group; and h) comparing the accumulated boundary values of step g against a Library of stored boundary values of known forms so as to determine a best match thereby identifying the unknown form.

Claim 9

A method for identifying business forms by constructing contour polygons from linear line segments that have been derived from a windowed raster scan of an image comprising the steps of: a) detecting the binary values of the pixels within the scanning window as the window is scanned over the image; b) positioning each of the binary values into an assigned and ordered position forming a string of binary values for each window scan with each string of binary values forming an integer number; c) directing the formed integer number to a look-up-table that translates each integer number into one or more output template commands; d) joining linear line segments into at least one contour polygon in response to the one or more output template commands; e) examining each contour polygon to determine if it represents a text component or a graphic component; f) grouping all text contour polygons and graphic contour polygons; g) accumulating the bounding values of the contour polygons of each group; and h) comparing the accumulated boundary values of step g against a Library of stored boundary values of known forms so as to determine a best match thereby identifying the unknown form.