Conversion of mechanical markings on a hardcopy document into machine-encoded annotations

1. A computer-implemented method of converting mechanical markings on hardcopy textual content to digital annotations in a digital document file, the method comprising:

identifying at least one block of text in a digital image of a hardcopy document;

converting textual content identified in the at least one block of text into machine-encoded text;

storing the machine-encoded text in a text block vector representing one line of text in the digital image of the hardcopy document;

identifying a type of at least one mechanical marking in the digital image of the hardcopy document, wherein the at least one mechanical marking covers or lies adjacent to the at least one block of text;

generating a machine-encoded annotated digital document file from the text block vector, the annotated digital document file being different from the digital image of the hardcopy document; and

modifying the text block vector in the annotated digital document file by marking the text block vector as the type of the at least one mechanical marking identified in the digital image of the hardcopy document, wherein the annotated digital document file includes the text block vector marked with computer-executable instructions representing at least one annotation corresponding to the at least one mechanical marking.

2. The method of claim 1 , wherein identifying the at least one block of text in the digital image of the hardcopy document comprises:

applying a Gaussian blur to the digital image of the hardcopy document to produce a blurred image of the at least one block of text;

performing adaptive thresholding on the blurred image to produce a segmented image of the at least one block of text;

applying a dilation operator to the segmented image to produce a morphed image of the at least one block of text; and

generating at least one bounding box surrounding one or more words in the morphed image of the at least one block of text.

3. The method of claim 1 , further comprising modifying the text block vector to include the computer-executable instructions representing the at least one annotation corresponding to the at least one mechanical marking identified in the digital image of the hardcopy document.

4. The method of claim 1 , wherein the type of the at least one mechanical marking is underlined text, colored/highlighted text, encircled text, or struckthrough text, and wherein identifying the type of the at least one mechanical marking in the digital image of the hardcopy document comprises:

a step for identifying the underlined text;

a step for identifying the colored/highlighted text;

a step for identifying the encircled text; or

a step for identifying the struckthrough text.

5. The method of claim 1 , wherein identifying the type of the at least one mechanical marking in the digital image of the hardcopy document comprises identifying the underlined text by locating one or more edges of the digital image of the hardcopy document to produce an edge image;

performing a Probabilistic Hough transformation on the edge image to produce a vector of lines;

marking the at least one block of text as underlined based on the vector of lines; and

converting the at least one block of text, with underlining, to machine-encoded text.

6. The method of claim 1 , wherein identifying the type of the at least one mechanical marking in the digital image of the hardcopy document comprises identifying the colored/highlighted text by converting at least a portion of the digital image of the hardcopy document containing the at least one block of text into a HSV (Hue, Saturation, Value) color space value;

storing the HSV color space value in a hue vector;

converting the HSV color space value into grayscale to produce a grayscale image of the converted portion of the digital image;

performing adaptive thresholding on the grayscale image to produce a segmented grayscale image;

marking the at least one block of text as colored based on the segmented grayscale image; and

converting the at least one block of text, with coloring, to machine-encoded text.

7. The method of claim 1 , wherein identifying the type of the at least one mechanical marking in the digital image of the hardcopy document comprises identifying the encircled text by

performing a Probabilistic Hough transformation on the at least one text block to produce a list of intersecting lines;

marking the at least one block of text as encircled based on the list of intersecting lines; and

converting the at least one block of text, with encircling, to machine-encoded text.

8. The method of claim 1 , wherein identifying the type of the at least one mechanical marking in the digital image of the hardcopy document comprises identifying the struckthrough text by

performing a Probabilistic Hough transformation on the at least one text block to produce a vector of lines;

marking the at least one block of text as struckthrough based on the vector of lines; and

converting the at least one block of text, with strike-through, to machine-encoded text.

9. The method of claim 1 , wherein the annotated digital document is compliant with a Portable Document Format (PDF) standard.

10. A non-transitory computer readable medium having instructions encoded thereon that when executed by one or more computer processors cause the one or more computer processors to perform a process comprising:

identifying at least one block of text in a digital image of a hardcopy document;

storing machine-encoded text, representing the at least one block of text, in a text block vector representing one line of text in the digital image of the hardcopy document;

identifying a type of at least one mechanical marking in the digital image of the hardcopy document, in which the at least one mechanical marking covers or lies adjacent to the at least one block of text;

generating a machine-encoded annotated digital document file from the text block vector, the annotated digital document file being different from the digital image of the hardcopy document; and

modifying the text block vector in the annotated digital document file by marking the text block vector as the type of the at least one mechanical marking identified in the digital image of the hardcopy document, wherein the annotated digital document file includes the text block vector marked with computer-executable instructions representing at least one annotation corresponding to the at least one mechanical marking.

11. The non-transitory computer readable medium of claim 10 , wherein identifying the at least one block of text in the digital image of the hardcopy document comprises:

applying a Gaussian blur to the digital image of the hardcopy document to produce a blurred image of the at least one block of text;

performing adaptive thresholding on the blurred image to produce a segmented image of the at least one block of text;

applying a dilation operator to the segmented image to produce a morphed image of the at least one block of text; and

generating at least one bounding box surrounding one or more words in the morphed image of the at least one block of text.

12. The non-transitory computer readable medium of claim 11 , further comprising modifying the text block vector to include the computer-executable instructions representing the at least one annotation corresponding to the at least one mechanical marking identified in the digital image of the hardcopy document.

13. The non-transitory computer readable medium of claim 10 , wherein the type of the at least one mechanical marking is underlined text, colored/highlighted text, encircled text, or struckthrough text, and wherein identifying the type of the at least one mechanical marking in the digital image of the hardcopy document comprises:

step for identifying the underlined text;

step for identifying the colored/highlighted text;

step for identifying the encircled text; or

step for identifying the struckthrough text.

14. The non-transitory computer readable medium of claim 10 , wherein identifying the type of the at least one mechanical marking in the digital image of the hardcopy document comprises identifying the underlined text by

locating one or more edges of the digital image of the hardcopy document to produce an edge image;

performing a Probabilistic Hough transformation on the edge image to produce a vector of lines;

marking the at least one block of text as underlined based on the vector of lines; and

converting the at least one block of text, with underlining, to machine-encoded text.

15. The non-transitory computer readable medium of claim 10 , wherein identifying the type of the at least one mechanical marking in the digital image of the hardcopy document comprises identifying the colored/highlighted text by

converting at least a portion of the digital image of the hardcopy document containing the at least one block of text into a HSV (Hue, Saturation, Value) color space value;

storing the HSV color space value in a hue vector;

converting the HSV color space value into grayscale to produce a grayscale image of the converted portion of the digital image;

performing adaptive thresholding on the grayscale image to produce a segmented grayscale image;

marking the at least one block of text as colored based on the segmented grayscale image; and

converting the at least one block of text, with coloring, to machine-encoded text.

16. The non-transitory computer readable medium of claim 10 , wherein identifying the type of the at least one mechanical marking in the digital image of the hardcopy document comprises identifying the encircled text by performing a Probabilistic Hough transformation on the at least one text block to produce a list of intersecting lines;

marking the at least one block of text as encircled based on the list of intersecting lines; and

converting the at least one block of text, with encircling, to machine-encoded text.

17. The non-transitory computer readable medium of claim 10 , wherein identifying the type of the at least one mechanical marking in the digital image of the hardcopy document comprises identifying the struckthrough text by

performing a Probabilistic Hough transformation on the at least one text block to produce a vector of lines;

marking the at least one block of text as struckthrough based on the vector of lines; and

converting the at least one block of text, with strike-through, to machine-encoded text.

18. The non-transitory computer readable medium of claim 10 , wherein the annotated digital document is compliant with a Portable Document Format (PDF) standard.

19. A system for converting mechanical markings on hardcopy textual content to digital annotations in a digital document file, the system comprising:

a storage; and

a processor operatively coupled to the storage, the processor configured to execute instructions stored in the storage that when executed cause the processor to carry out a process including

identifying at least one block of text in a digital image of a hardcopy document

storing machine-encoded text, representing at least one line of text in a digital image of a hardcopy document, in a text block vector;

identifying a type of at least one mechanical marking in the digital image of the hardcopy document, in which the at least one mechanical marking covers or lies adjacent to the at least one block of text;

generating a machine-encoded annotated digital document file from text block vector, the annotated digital document file being different from the digital image of the hardcopy document; and

modifying the text block vector in the annotated digital document file by marking the text block vector as the type of the at least one mechanical marking identified in the digital image of the hardcopy document, wherein the annotated digital document file includes the text block vector marked with computer-executable instructions representing at least one annotation corresponding to the at least one mechanical marking.

20. The system of claim 19 , wherein the type of the at least one mechanical marking is underlined text, colored/highlighted text, encircled text, or struckthrough text, and wherein identifying the type of the at least one mechanical marking in the digital image of the hardcopy document comprises:

identifying the underlined text by

locating one or more edges of the digital image of the hardcopy document to produce an edge image,

performing a Probabilistic Hough transformation on the edge image to produce a vector of lines,

marking the at least one block of text as underlined based on the vector of lines, and

converting the at least one block of text, with underlining, to machine-encoded text;

identifying the colored/highlighted text by

converting at least a portion of the digital image of the hardcopy document containing the at least one block of text into a HSV (Hue, Saturation, Value) color space value,

storing the HSV color space value in a hue vector,

converting the HSV color space value into grayscale to produce a grayscale image of the converted portion of the digital image,

performing adaptive thresholding on the grayscale image to produce a segmented grayscale image,

marking the at least one block of text as colored based on the segmented grayscale image, and

converting the at least one block of text, with coloring, to machine-encoded text;

identifying the encircled text by

performing a Probabilistic Hough transformation on the at least one text block to produce a list of intersecting lines,

marking the at least one block of text as encircled based on the list of intersecting lines, and

converting the at least one block of text, with encircling, to machine-encoded text; and

identifying the struckthrough text by

performing a Probabilistic Hough transformation on the at least one text block to produce a vector of lines,

marking the at least one block of text as struckthrough based on the vector of lines, and

converting the at least one block of text, with strike-through, to machine-encoded text.