IP Library Granted Patent US 12,430,724
Granted Patent B2
US 12,430,724 · App. 17/483,295 · Granted Sep 30, 2025

Autonomously removing scan marks from digital documents utilizing content-aware filters

Inventors: Aishwarya Mittal (Mathura, IN); Sachin Beniwal (Hisar, IN); Abhishek Kumar Pandey (Katihar, IN)
Assignee: Adobe Inc.
G06T5/77G06T5/60G06T7/11G06T7/136G06T2207/30176
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,430,724
App. No.
17/483,295
Granted
Sep 30, 2025
Kind
B2
Abstract

The present disclosure relates to systems, non-transitory computer-readable media, and methods for implementing content-aware filters to autonomously remove scan marks from digital documents. In particular implementations, the disclosed systems utilize a set of targeted scan mark models in a scan mark removal pipeline. For example, each scan mark model includes a corresponding content-aware filter configured to identify document regions that match a designated class of scan marks to filter. Examples of scan mark models include staple scan mark models, punch hole scan mark models, and page turn scan mark models. In certain embodiments, the disclosed systems then use the scan mark models to generate mark-specific masks based on document input features. Additionally, in some embodiments, the disclosed systems combine the mark-specific masks into a final segmentation mask and apply the final segmentation mask to the digital document for correcting the identified regions with scan marks.

Claims (58)

1. A non-transitory computer-readable medium comprising instructions that, when executed by at least one processor, cause a computing device to:

generate, from a digital document comprising one or more scan marks, a margin boundary, an OCR mask comprising a binary image having regions for detected text, and digital contours defining regions within the digital document;

generate, utilizing a staple scan mark model, a staple mask from the margin boundary, the OCR mask, and the digital contours by applying a threshold region area for staples and a threshold region width and height for staples;

generate, utilizing a punch hole scan mark model, a punch hole mask from the margin boundary, the OCR mask, and the digital contours by applying a threshold region area for punch holes and a threshold region width and height for punch holes; and

generate a modified digital document without the one or more scan marks by removing a staple region corresponding to the staple mask and a punch hole region corresponding to the punch hole mask from the digital document.

2. The non-transitory computer-readable medium of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the computing device to generate the staple mask from the digital contours by applying the threshold region area for staples by filtering out a region that has digital contours satisfying a maximum area value corresponding to staple scan marks.

3. The non-transitory computer-readable medium of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the computing device to generate the staple mask from the digital contours by applying the threshold region width and height for staples by filtering out a region that has digital contours satisfying maximum width and height values corresponding to staple scan marks.

4. The non-transitory computer-readable medium of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the computing device to generate the punch hole mask from the digital contours by:

applying the threshold region area for punch holes by filtering out a first region that has digital contours satisfying a maximum area value corresponding to punch hole scan marks; and

applying the threshold region width and height for punch holes by filtering out a second region that has digital contours satisfying maximum width and height values corresponding to punch hole scan marks.

5. The non-transitory computer-readable medium of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the computing device to:

generate, utilizing a page turn scan mark model, a page turn mask from the margin boundary, the OCR mask, and the digital contours by applying a threshold region area for page turns and a threshold region width and height for page turns; and

generate the modified digital document without the one or more scan marks further by removing a page turn region corresponding to the page turn mask from the digital document.

6. The non-transitory computer-readable medium of claim 5 , further comprising instructions that, when executed by the at least one processor, cause the computing device to:

generate a final segmentation mask by combining the staple mask, the punch hole mask, and the page turn mask; and

remove the one or more scan marks utilizing the final segmentation mask.

7. The non-transitory computer-readable medium of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the computing device to validate the staple region and the punch hole region by:

excluding, from a subset of regions comprising the staple region and the punch hole region, one or more regions by comparing a position of the one or more regions with the margin boundary; and

excluding, from the subset of regions, at least one region by comparing a position of the at least one region with the OCR mask.

8. The non-transitory computer-readable medium of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the computing device to:

generate the margin boundary based on a page size for the digital document and a bounding box corresponding to document content of the digital document;

generate the OCR mask utilizing optical character recognition; and

generate the digital contours based on polygonal approximations of the regions within the digital document.

9. A system comprising:

one or more memory devices; and

one or more processors coupled to the one or more memory devices that cause the system to perform operations comprising:

generating, from a digital document comprising one or more scan marks, a margin boundary, an OCR mask comprising a binary image having regions for detected text, and digital contours defining regions within the digital document;

generating, utilizing a staple scan mark model, a staple mask from the margin boundary, the OCR mask, and the digital contours by applying a threshold region area for staples and a threshold region width and height for staples;

generating, utilizing a punch hole scan mark model, a punch hole mask from the margin boundary, the OCR mask, and the digital contours by applying a threshold region area for punch holes and a threshold region width and height for punch holes; and

generating a modified digital document without the one or more scan marks by removing a staple region corresponding to the staple mask and a punch hole region corresponding to the punch hole mask from the digital document.

10. The system of claim 9 , wherein the operations further comprise generating the staple mask from the digital contours by applying the threshold region area for staples by filtering out a region that has digital contours satisfying a maximum area value corresponding to staple scan marks.

11. The system of claim 9 , wherein the operations further comprise generating the staple mask from the digital contours by applying the threshold region width and height for staples by filtering out a region that has digital contours satisfying maximum width and height values corresponding to staple scan marks.

12. The system of claim 9 , wherein the operations further comprise generating the punch hole mask from the digital contours by:

applying the threshold region area for punch holes by filtering out a first region that has digital contours satisfying a maximum area value corresponding to punch hole scan marks; and

applying the threshold region width and height for punch holes by filtering out a second region that has digital contours satisfying maximum width and height values corresponding to punch hole scan marks.

13. The system of claim 9 , wherein the operations further comprise:

generating, utilizing a page turn scan mark model, a page turn mask from the margin boundary, the OCR mask, and the digital contours by applying a threshold region area for page turns and a threshold region width and height for page turns; and

generating the modified digital document without the one or more scan marks further by removing a page turn region corresponding to the page turn mask from the digital document.

14. The system of claim 13 , wherein the operations further comprise:

generating a final segmentation mask by combining the staple mask, the punch hole mask, and the page turn mask; and

removing the one or more scan marks utilizing the final segmentation mask.

15. A computer-implemented method comprising:

generating, from a digital document comprising one or more scan marks, a margin boundary, an OCR mask comprising a binary image having regions for detected text, and digital contours defining regions within the digital document;

generating, utilizing a staple scan mark model, a staple mask from the margin boundary, the OCR mask, and the digital contours by applying a threshold region area for staples and a threshold region width and height for staples;

generating, utilizing a punch hole scan mark model, a punch hole mask from the margin boundary, the OCR mask, and the digital contours by applying a threshold region area for punch holes and a threshold region width and height for punch holes; and

generating a modified digital document without the one or more scan marks by removing a staple region corresponding to the staple mask and a punch hole region corresponding to the punch hole mask from the digital document.

16. The computer-implemented method of claim 15 , further comprising generating the staple mask from the digital contours by applying the threshold region area for staples by filtering out a region that has digital contours satisfying a maximum area value corresponding to staple scan marks.

17. The computer-implemented method of claim 15 , further comprising generating the staple mask from the digital contours by applying the threshold region width and height for staples by filtering out a region that has digital contours satisfying maximum width and height values corresponding to staple scan marks.

18. The computer-implemented method of claim 15 , further comprising generating the punch hole mask from the digital contours by:

applying the threshold region area for punch holes by filtering out a first region that has digital contours satisfying a maximum area value corresponding to punch hole scan marks; and

applying the threshold region width and height for punch holes by filtering out a second region that has digital contours satisfying maximum width and height values corresponding to punch hole scan marks.

19. The computer-implemented method of claim 15 , further comprising validating the staple region and the punch hole region by:

excluding, from a subset of regions comprising the staple region and the punch hole region, one or more regions by comparing a position of the one or more regions with the margin boundary; and

excluding, from the subset of regions, at least one region by comparing a position of the at least one region with the OCR mask.

20. The computer-implemented method of claim 15 , further comprising:

generating the margin boundary based on a page size for the digital document and a bounding box corresponding to document content of the digital document;

generating the OCR mask utilizing optical character recognition; and

generating the digital contours based on polygonal approximations of the regions within the digital document.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Sep 23, 2021
From: MITTAL, AISHWARYA; BENIWAL, SACHIN; PANDEY, ABHISHEK KUMAR
To: ADOBE INC.
Reel/Frame 057581/0028 →
Continuity (1)
Related Publication 20230090313A1 · Mar 23, 2023
References Cited (10)
US 5448376A · Ohta · 1995 [cited by examiner]
US 20050128516A1 · Tomita · 2005 [cited by examiner]
US 20090086275A1 · Liang · 2009 [cited by examiner]
US 20210312607A1 · Lin · 2021 [cited by examiner]
CN 111754414A · 2020 [cited by examiner]
JP 2006060520A · 2006 [cited by examiner]
JP 2011024116A · 2011 [cited by examiner]
Lins, “A Taxonomy for Noise in Images of Paper Documents—The Physical Noises,” International Conference Image Analysis and Recognition (ICIAR), Lecture Notes in Computer Science (LNCP), vol. 5627, pp. 844-854 (Year: 200… [cited by examiner]
A. Kolesnikov and P. Fränti, Reduced-Search Dynamic Programming For Approximation Of Polygonal Curves, In Pattern Recognition Letters, vol. 24 (14), pp. 2243-2254, 2003. [cited by applicant]
U.S. Appl. No. 17/202,019, entitled Generating Modified Digital Images Using Deep Visual Guided Patch Match Models for Image Inpainting; filed on Mar. 15, 2021. [cited by applicant]