Method and apparatus for data capture from imaged documents

• US 7,911,660 B2
• Filed: 06/18/2008
• Issued: 03/22/2011
• Est. Priority Date: 08/11/2000
• Status: Expired due to Term

• Current Assignee: McGraw Hill LLC
• Original Assignee: CTB/McGraw Hill LLC
• Inventors: Poor, David D. S.
• Primary Examiner(s): Worku; Negussie
• Application Number:
  US12/141,349
• Publication Number:
  US 20080316552A1
• Time in Patent Office: 1,007 Days
• Field of Search: 358/474, 358/501, 358/1.2, 358/1.14, 358/1.9, 358/518, 382/197, 382/188, 382/203, 382/287, 382/291, 382/317, 235/456, 235/495, 235/454, 434/322, 434/353, 434/362, 434/323
• US Class Current: 358/474
• CPC Class Codes: G06K 17/0032 Apparatus for automatic testing and analysing marked record carriers, used for examinations of the multiple choice answer type teaching apparatus working with questions and answers G09B7/00
• CPC Class Codes: G06K 7/12 using a selected wavelength, e.g. to sense red marks and ignore blue marks
• CPC Class Codes: G06K 7/1447 extracting optical codes from image or text carrying said optical code
• CPC Class Codes: G06V 10/143 Sensing or illuminating at different wavelengths
• CPC Class Codes: G06V 10/22 by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
• CPC Class Codes: G06V 10/242 by image rotation, e.g. by 90 degrees
• CPC Class Codes: G06V 10/25 Determination of region of interest [ROI] or a volume of interest [VOI]

Abstract

Methods and apparatus are disclosed for locating an area of interest within a digital image of a form captured by an imaging scanner. Specific examples include methods and apparatus for optical mark reading with a digital imaging scanner. In many of the methods, an image of a response form is captured by a scanner, and “target” areas for possible responses are located within the image based upon an expected location being adjusted as necessary for certain error-inducing defects in the forms or scanning process. Also disclosed are steps to normalize the darkness values of pixels captured from an optically scanned form.

Claims

1. A computer-implemented method of normalizing darkness values of pixels captured from a form by an optical scanning device, to enable subsequent processing of the contents of the form based on a relative darkness of the pixels, said method comprising the steps of:

a) scanning the form with an optical scanning device capable of producing pixel darkness values related to a relative amount of light reflected from the form;

b) recording on computer-readable media a darkness value of individual pixels from the scanned form;

c) defining an area of the form to examine for determining pixel normalization parameters;

d) identifying the pixel darkness value corresponding to a least amount of light reflected within the area;

e) determining a normalization scale such that the pixel darkness value corresponding to the least amount of light reflected corresponds to a darkness value corresponding approximately to pure black; and

f) applying the normalization scale to the recorded darkness values to create normalized darkness values.

2. A method as in claim 1, wherein the optical scanning device is capable of producing pixel darkness values corresponding to a standard scale of darkness increments between pure black and pure white, and the step of determining a normalization scale is accomplished by the sub-steps of:

a) providing a plurality of look-up tables each having a differentiated scale of fewer increments than the standard scale, and

b) selecting a look-up table based upon a difference between the pixel darkness value corresponding to the least amount of reflected light and the standard scale value corresponding to pure black.

3. A method as in claim 2 wherein the standard scale has increments from 0-255 from pure black to pure white.

4. A computer-implemented method of normalizing darkness values of pixels captured from a form by an optical scanning device, to enable subsequent processing of the contents of the form based on a relative darkness of the pixels, said method comprising the steps of:

a) scanning the form with an optical scanning device capable of producing pixel darkness values related to a relative amount of light reflected from the form;

b) recording on computer-readable media a darkness value of individual pixels from the scanned form;

c) defining an area of the form to examine for determining pixel normalization parameters;

d) identifying the pixel darkness value corresponding to a greatest amount of light reflected within the area;

e) determining a normalization scale such that the pixel darkness value corresponding to the greatest amount of light reflected corresponds to a darkness value corresponding approximately to pure white; and

f) applying the normalization scale to the recorded darkness values to create normalized darkness values.

5. A method as in claim 4, wherein the optical scanning device is capable of producing pixel darkness values corresponding to a standard scale of darkness increments between pure black and pure white, and the step of determining a normalization scale is accomplished by the sub-steps of:

a) providing a plurality of look-up tables each having a differentiated scale of fewer increments than the standard scale, and

b) selecting a look-up table based upon a difference between the pixel darkness value corresponding to the greatest amount of reflected light and the standard scale value corresponding to pure white.

6. A method as in claim 5, wherein the standard scale has increments from 0-255 from pure black to pure white.

7. A computer-implemented method of normalizing darkness values of pixels captured from a form by an optical scanning device, to enable subsequent processing of the contents of the form based on a relative darkness of the pixels, said method comprising the steps of:

a) scanning the form with an optical scanning device capable of producing pixel darkness values related to a relative amount of light reflected from the form;

b) recording on computer-readable media the darkness value of individual pixels from the scanned form;

c) defining an area of the form to examine for determining pixel normalization parameters;

d) identifying the pixel darkness values corresponding to a greatest amount of light reflected and a least amount of light reflected within the area;

e) determining a normalization scale such that the pixel darkness value corresponding to the greatest amount of light reflected corresponds to a darkness value corresponding approximately to pure white and the pixel darkness value corresponding to the least amount of light reflected corresponds to a darkness value corresponding approximately to pure black; and

f) applying the normalization scale to the recorded darkness values to create normalized darkness values.

8. A computer-implemented method of normalizing darkness values of pixels captured from a form by an optical scanning device having multiple optical elements, to enable subsequent processing of the contents of the form based on a relative darkness of the pixels, said method comprising the steps of:

a) scanning the form with an optical scanning device having multiple optical elements capable of producing pixel darkness values related to a relative amount of light reflected off the form to an element;

b) recording on computer-readable media the pixel darkness values produced by each optical element;

c) defining a section of the form to examine for determining pixel normalization parameters;

d) identifying, for each optical element, the pixel darkness value corresponding to a least amount of light reflected within the section;

e) determining a normalization scale for each optical element such that the pixel darkness value corresponding to the least amount of light reflected corresponds to a darkness value corresponding approximately to pure black; and

f) applying the appropriate normalization scale to the recorded darkness values to create normalized darkness values.

9. A method as in claim 8, wherein each optical scanning element is capable of producing pixel darkness values corresponding to a standard scale of darkness increments between pure black and pure white, and the step of determining a normalization scale is accomplished by the sub-steps of:

a) providing a plurality of look-up tables each having a differentiated scale of fewer increments than the standard scale, and

b) selecting a look-up table based upon a difference between the pixel darkness value corresponding to the least amount of reflected light and the standard scale value corresponding to pure black.

10. A computer-implemented method of normalizing darkness values of pixels captured from a form by an optical scanning device having multiple optical elements, to enable subsequent processing of the contents of the form based on a relative darkness of the pixels, said method comprising the steps of:

a) scanning the form with an optical scanning device having multiple optical elements capable of producing pixel darkness values related to a relative amount of light reflected off the form to an element;

b) recording on computer-readable media the pixel darkness values produced by each optical element;

c) defining a section of the form to examine for determining pixel normalization parameters;

d) identifying, for each optical element, the pixel darkness value corresponding to a greatest amount of light reflected within the section;

e) determining a normalization scale for each optical element such that the pixel darkness value corresponding to the greatest amount of light reflected corresponds to a darkness value corresponding approximately to pure white; and

f) applying the appropriate normalization scale to the recorded darkness values to create normalized darkness values.

11. A method as in claim 10, wherein each optical scanning element is capable of producing pixel darkness values corresponding to a standard scale of darkness increments between pure black and pure white, and the step of determining a normalization scale is accomplished by the sub-steps of:

a) providing a plurality of look-up tables each having a differentiated scale of fewer increments than the standard scale, and

b) selecting a look-up table based upon a difference between the pixel darkness value corresponding to the greatest amount of reflected light and the standard scale value corresponding to pure white.

12. A computer-implemented method of normalizing darkness values of pixels captured from a form by an optical scanning device having multiple optical elements, to enable subsequent processing of the contents of the form based on a relative darkness of the pixels, said method comprising the steps of:

a) scanning the form with an optical scanning device having multiple optical elements capable of producing pixel darkness values related to a relative amount of light reflected off the form to an element;

b) recording on computer-readable media the pixel darkness values produced by each optical element;

c) defining a section of the form to examine for determining pixel normalization parameters;

d) identifying, for each optical element, the pixel darkness value corresponding to a greatest amount of light reflected to any optical element within the section;

e) identifying the pixel darkness value corresponding to a least amount of light reflected to any optical element within the section;

f) determining a normalization scale for each optical element such that the pixel darkness value corresponding to the greatest amount of light reflected corresponds to a darkness value corresponding approximately to pure white and the pixel darkness value corresponding to the least amount of light corresponds to approximately pure black; and

g) applying the appropriate normalization scale to the recorded darkness values to create normalized darkness values.

13. A computer-implemented method of normalizing the darkness value of pixels captured by optically scanning a form, to enable subsequent processing of the contents of the form based on a relative darkness of the pixels, said method comprising the steps of:

a) scanning the form with an optical imaging scanner having an array of optical elements capable of producing an image of the form and, for each pixel of the image, a gray scale value of light reflected off the form, the scale ranging from 0 for pure black to 255 for pure white;

b) recording on computer-readable media the gray scale value for each pixel;

c) defining an area within the image of the form to examine for determining pixel normalization parameters;

d) detecting, for each optical element, a highest recorded gray scale value within the area;

e) determining a normalization scale for each optical element based upon a difference between its highest gray scale value and 255; and

f) applying the associated normalization scale to pixels within the captured image to covert gray scale values to normalized values.

14. A computer-implemented method of normalizing the darkness value of pixels captured by optically scanning a form, to enable subsequent processing of the contents of the form based on a relative darkness of the pixels, said method comprising the steps of:

a) scanning the form with an optical imaging scanner having an array of optical element capable of producing an image of the form and, for each pixel of the image, a gray scale value ranging from 0 for pure black to 255 for pure white;

b) recording on computer-readable media the gray scale value for each pixel;

c) defining an area within the image of the form to examine for determining pixel normalization parameters;

d) detecting, for each optical element, a highest recorded gray scale value within the area;

e) determining a white normalization scale for each optical element related to a difference between the highest gray scale value and 255;

f) applying the associated white normalization scale to pixels within the captured image;

g) detecting a lowest gray scale value recorded within the area;

h) determining a black normalization scale for each optical element related to a difference between the lowest gray scale value and 0; and

i) applying the black normalization scale to the white normalized values.

15. A computer-implemented method of normalizing a color intensity value of pixels captured from a form, to enable subsequent processing of the contents of the form based on a relative darkness of the pixels, said method comprising the steps of:

a) scanning the form with an optical imaging scanner capable of producing an image of the form and, for each pixel in the image, a color scale value related to a relative amount of light in a particular color band that is reflected from the form;

b) recording on computer-readable media the color scale value for each pixel in the image;

c) defining an area within the image of the form to examine for pixel normalization;

d) detecting a highest color scale value of pixels within the area;

e) determining a normalization parameter related to the difference between the highest color scale value of pixels within the area to the highest value of the scale; and

f) applying the normalization parameter to pixels within the captured image to covert a pixel'"'"'s color scale value to a normalized value.

16. A computer-implemented method of normalizing a color scale value of pixels captured from a form, to enable subsequent processing of the contents of the form based on a relative darkness of the pixels, said method comprising the steps of:

a) scanning the form with an optical imaging scanner capable of producing an image of the form and, for each pixel in the image, a color scale value related to a relative amount of light in a particular color band that is reflected by the form;

b) recording on computer-readable media a color scale value for each pixel in the image;

c) defining an area within the image of the form to examine for determining pixel normalization parameters;

d) detecting a highest and a lowest color scale values of pixels within the area;

e) determining a normalization parameter related to a difference between the highest color scale value of pixels within the area to the highest value of the scale, and related to a difference between the lowest color scale value of pixels within the area to the lowest value of the scale; and

f) applying the normalization parameter to pixels within the captured image to covert a pixel'"'"'s color scale value to a normalized value.

17. A computer-implemented method of converting a gray scale scanned image to a bi-tonal digital image, the method comprising the steps of:

a) scanning a form with an optical imaging scanner capable of producing an image of the form and, for each pixel in the image, a gray scale value related to a relative amount of light reflected by the form;

b) recording on computer-readable media a gray scale value for each pixel in the image;

c) defining an area within the image of the form to examine for pixel normalization;

d) detecting a highest gray scale value of pixels within the area;

e) determining a normalization parameter related to the difference between the highest gray scale value of pixels within the area to a value corresponding approximately to a pure white sheet;

f) applying the normalization parameter to pixels within the captured image of the form to covert a pixel'"'"'s gray scale value to a normalized value;

g) selecting a cut-off value within the gray scale; and

h) converting all pixels with normalized gray values above the cut-off value to the bi-tonal digit 0 (white), and all pixels with normalized gray values below the cut-off value to the bi-tonal digit 1 (black).

18. A computer-implemented method of optical mark reading (OMR) with an optical imaging scanner, said method comprising the steps of:

a) scanning a sheet printed with a plurality of OMR response targets with a scanner capable of producing an image of the sheet and any OMR response marks made thereon, and, for each pixel in the image, a scale value related to a relative amount of light reflected from the sheet;

b) recording on computer-readable media the scale value of reflected light for each pixel;

c) defining an area within the image to examine for pixel normalization;

d) detecting a highest scale value of pixels within the area;

e) determining a normalization parameter related to the difference between the highest scale value of pixels within the area to a value corresponding approximately to a pure white sheet;

f) applying the normalization parameter to pixels within the captured image of the form to covert a pixel'"'"'s scale value to a normalized value;

g) selecting a cut-off value within the scale;

h) converting all pixels with normalized values above the cut-off to the bi-tonal digit 0 (white) and all pixels with normalized values below the cut-off to the bi-tonal digit 1 (black); and

i) interrogating pixels within a proscribed area around an expected location of an OMR response target for bi-tonal values of 1 as an indication of an OMR response mark.