Segmentation of handwriting and machine printed text

• US 5,181,255 A
• Filed: 12/13/1990
• Issued: 01/19/1993
• Est. Priority Date: 12/13/1990
• Status: Expired due to Term

• Current Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Bloomberg, Dan S.
• Primary Examiner(s): Moore, David K.
• Assistant Examiner(s): Stellrecht, Barry
• Application Number:
  US07/627,284
• Time in Patent Office: 768 Days
• Field of Search: 382/9, 382/36, 382/48, 382/55, 382/7
• US Class Current: 382/176
• CPC Class Codes: G06V 30/2455 Discrimination between machine-print, hand-print and cursive writing
• CPC Class Codes: G06V 30/413 Classification of content, e.g. text, photographs or tables

Abstract

A method and apparatus for differentiating and extracting handwritten annotations and machine printed text in an image. The method provides for the use of morphological operations, preferably at reduced scale, to eliminate for example, the handwritten annotations from an image. A separation mask is produced that, for example, covers all the image pixels corresponding to machine printed text, and none of the image pixels corresponding to handwritten or handprinted annotations. The separation mask is used in conjunction with the original image to produce separate handwritten annotations and machine printed text images.

Claims

1. In a digital processing means, a method of identifying handwritten annotation areas of an input image having handwritten annotation areas and machine printed text areas comprising morphologically processing a region of said input image having a plurality of handwritten and machine printed characters to produce a destination image, said destination image identifying substantially only said machine printed text or said handwritten annotations areas, wherein said step of processing further comprises the step of identifying regions of machine printed text or handwritten annotations with a structuring element (SE) which selectively identifies said machine printed text or handwritten annotations to produce a first intermediate image, wherein the step of identifying regions further comprises the steps of:

a) closing said first intermediate image with an SE having two horizontally adjacent ON pixels to produce a second intermediate image; and

b) xoring said input image with said second intermediate image.

2. In a digital processing means, a method of identifying handwritten annotation areas of an input image having handwritten annotation areas and machine printed text areas comprising morphologically processing a region of said input image having a plurality of handwritten and machine printed characters to produce a destination image, said destination image identifying substantially only said machine printed text or said handwritten annotations areas, wherein said step of processing further comprises the step of identifying regions of machine printed text or handwritten annotations with a structuring element (SE) which selectively identifies said machine printed text or handwritten annotations to produce a first intermediate image, and further comprising the steps of:

a) closing said first intermediate image with a SE having at least two vertically adjacent ON pixels;

b) openings said first intermediate image with a SE having at least two vertically adjacent ON pixels.

3. In a digital processing means, a method of identifying handwritten annotation areas of an input image having handwritten annotation areas and machine printed text areas comprising morphologically processing a region of said input image having a plurality of handwritten and machine printed characters to produce a destination image, said destination image identifying substantially only said machine printed text or said handwritten annotations areas, wherein said step of processing further comprises the step of identifying regions of machine printed text or handwritten annotations with a structuring element (SE) which selectively identifies said machine printed text or handwritten annotations to produce a first intermediate image, and further comprising the step of forming a mask from said first intermediate image by closing and opening said image with SEs having a plurality of laterally adjacent ON pixels to produce a second intermediate image.

4. The method as recited in claim 3 further comprising the steps of:

a) performing a thresholded reduction of said second intermediate image;

b) opening the output of step a) with a horizontal SE;

c) dilating the output of step b) with a horizontal SE;

d) closing the output of step c) with a vertical SE;

e) dilating the output of step d) with a horizontal SE; and

f) dilating the output of step e) with a vertical SE.

5. In a digital processing means, a method of identifying handwritten annotation areas of an input image having handwritten annotation areas and machine printed text areas comprising morphologically processing a region of said input image having a plurality of handwritten and machine printed characters to produce a destination image, said destination image identifying substantially only said machine printed text or said handwritten annotations areas, wherein said destination image is a mask, said mask overlaying substantially only said machine printed text areas or handwritten annotation areas, wherein said mask is formed by the steps of:

a) extracting seed pixels in substantially only handwritten annotation or machine printed text areas; and

b) filling said seed pixels to a clipping mask, said clipping mask comprising substantially solid regions of ON pixels over said handwritten annotation areas or said machine printed text areas, further comprising the step of forming said clipping mask, the step of forming said clipping mask further comprises the steps of;

i) closing output of step a) with a horizontal SE; and

ii) opening output of step b) with a horizontal SE.

6. The method as recited in claim 5 preceded by the step of reducing said input image.

7. The method as recited in claim 5 wherein said close and said open use horizontal SEs with more than three adjacent ON pixels.

8. In a digital processing means, a method of identifying handwritten annotation areas of an input image having handwritten annotation areas and machine printed text areas comprising morphologically processing a region of said input image having a plurality of handwritten and machine printed characters to produce a destination image, said destination image identifying substantially only said machine printed text or said handwritten annotations areas, wherein said destination image is a mask overlaying substantially only said machine printed text areas or said handwritten annotation areas, wherein said mask is formed by the steps of:

extracting seed pixels in substantially only said handwritten annotation or said machine printed text areas by identifying regions of said image having OFF pixels laterally bounded on each side by ON pixels; and

filling said seed pixels to a clipping mask, said clipping mask comprising substantially solid regions of ON pixels over said handwritten annotation areas and said machine printed text areas.

9. The method as recited in claim 8 wherein said closing step is preceded by the step of reducing said input image.

10. In a digital processing means, a method of identifying handwritten annotation areas of an input image having handwritten annotation areas and machine printed text areas comprising morphologically processing a region of said input image having a plurality of handwritten and machine printed characters to produce a destination image, said destination image identifying substantially only said machine printed text or said handwritten annotations areas, wherein said destination image is a mask overlaying substantially only said machine printed text areas or said handwritten annotation areas, wherein said mask is formed by the steps of:

extracting seed pixels in substantially only said handwritten annotation or said machine printed text areas by

a) closing said input image; and

b) xoring output of step a) with said input image; and

filling said seed pixels to a clipping mask, said clipping mask comprising substantially solid regions of ON pixels over said handwritten annotation areas and said machine printed text areas.

11. In a digital processing means, a method of identifying handwritten annotation areas of an input image having handwritten annotation areas and machine printed text areas comprising morphologically processing a region of said input image having a plurality of handwritten and machine printed characters to produce a destination image, said destination image identifying substantially only said machine printed text or said handwritten annotations areas, wherein said destination image is a mask overlaying substantially only said machine printed text areas or said handwritten annotation areas, wherein said mask is formed by the steps of:

extracting seed pixels in substantially only said handwritten annotation or said machine printed text areas by processing said image with a hit-miss SE; and

filling said seed pixels to a clipping mask, said clipping mask comprising substantially solid regions of ON pixels over said handwritten annotation areas and said machine printed text areas.

12. In a digital processing means, a method of identifying handwritten annotation areas of an input image having handwritten annotation areas and machine printed text areas comprising morphologically processing a region of said input image having a plurality of handwritten and machine printed characters to produce a destination image, said destination image identifying substantially only said machine printed text or said handwritten annotations areas, wherein said destination image is a mask overlaying substantially only said machine printed text areas or said handwritten annotation areas, wherein said mask is formed by the steps of:

extracting seed pixels in substantially only said handwritten annotation or said machine printed text areas; and

filling said seed pixels to a clipping mask, said clipping mask comprising substantially solid regions of ON pixels over said handwritten annotation areas and said machine printed text areas, by the steps of;

a) storing an input image;

b) dilating said input image;

c) anding output from step b) with said clipping mask;

d) comparing output of step c) with a stored image from step a); and

e) if said stored image from step a) and said output of step c) differ by more than a predetermined amount, outputting said output of step c) as a mask or, if said stored image from step a) and said output of step c) differ by more than a predetermined amount, repeating to step a) with said output of step c) acting as said input image.