AUTOMATIC CAPTURING OF DOCUMENTS HAVING PRELIMINARILY SPECIFIED GEOMETRIC PROPORTIONS

• US 20140376060A1
• Filed: 01/23/2014
• Published: 12/25/2014
• Est. Priority Date: 06/19/2013
• Status: Active Grant

• Current Assignee: ABBYY Development LLC
• Original Assignee: ABBYY Development LLC
• Inventors: Loginov, Vasily, Zagaynov, Ivan, Bocharov, Konstantin, Lyubimov, lakov
• Granted Patent: US 9,071,745 B2
• Time in Patent Office: Days
• US Class Current: 358/474
• CPC Class Codes: H04N 1/00466 displaying finishing information, e.g. position of punch holes or staple or orientation references
• CPC Class Codes: H04N 1/00748 Detecting edges, e.g. of a stationary sheet
• CPC Class Codes: H04N 1/00771 Indicating or reporting, e.g. issuing an alarm
• CPC Class Codes: H04N 1/00816 Determining the reading area, e.g. eliminating reading of margins
• CPC Class Codes: H04N 1/00824 for displaying or indicating, e.g. a condition or state details of displaying or indicating means H04N1/00405
• CPC Class Codes: H04N 1/1072 Means for guiding the scanning, e.g. rules H04N1/1077 takes precedence
• CPC Class Codes: H04N 1/19594 using a television camera or a still video camera
• CPC Class Codes: H04N 1/23 Reproducing arrangements

Abstract

Methods, devices, and systems for capturing documents are provided. A frame in a viewfinder is selected and is converted into a binary version. Segmentation of the binarized version of the frame is performed using computed horizontal and vertical histograms. At least one document is identified in the selected frame. The at least one document is captured upon determining that values of the one or more parameters associated with the document match predetermined values of corresponding one or more parameters.

Claims

1. A method for automatically capturing of an electronic image of documents based on results of a preliminary analysis of a current view in a viewfinder, the method comprising:

selecting a frame in the viewfinder;

converting the frame into a binary version of the frame;

computing a horizontal histogram of distribution of pixels along horizontal direction of the binarized version of the frame;

computing a vertical histogram of distribution of pixels along vertical direction of the binarized version of the frame;

performing segmentation of the binarized version of the frame using the horizontal histogram and the vertical histogram resulting in one or more horizontal and vertical segments, wherein each of the one or more horizontal and vertical segments corresponds to one document included in the frame;

identifying at least one document in the selected frame, wherein one or more parameters are associated with the at least one document;

comparing values of one or more parameters associated with the at least one identified document with predetermined values for the one or more parameters; and

capturing the at least one document upon determining that values of the one or more parameters associated with the document match predetermined values corresponding to the one or parameters.

2. The method of claim 1, wherein the frame selection occurs at a predetermined frequency.

3. The method of claim 2, wherein the predetermined frequency specifies a number of frames per second.

4. The method of claim 1, wherein the predetermined values of the one or more predetermined parameters are received from a user interface allowing a user to specify values for the parameters.

5. The method of claim 1, wherein the one or more parameters include geometrical proportions.

6. The method of claim 1, wherein the one or more parameters include a parameter specifying a width to length ratio.

7. The method of claim 6, wherein the width to length ratio of the document matches the predetermined width to length ratio.

8. The method of claim 1, further comprising:

compressing the frame;

smoothing noise interferences in one or more search areas of the compressed frame;

detecting edges within each of the search areas of the compressed frame using a differential edge detection algorithm, wherein the detected edges form the document within the compressed frame; and

performing one or more checks on the detected edges within the search areas.

9. The method of claim 8, wherein performing one or more checks on the detected edges within the search areas is based on checking at least one of the reliability criteria to determine baseline for one or more candidate documents, and wherein the process of identifying baselines is performed in an iterative manner based on searching for additional detected edges, until additional detected edges will meet the reliability criteria.

10. The method of claim 8, wherein detecting edges in the detected current view includes applying techniques of suppression non-maximum pixel values within regions of interest of the detected current view by applying one or more threshold values.

11. The method of claim 8, wherein detecting edges in the detected current view includes applying a techniques of searching edges with maximum value of intensity of a contour, wherein searching edges with maximum value of intensity of a contour is performed based on Hough transform and exhaustive search among all possible candidates.

12. The method of claim 1, wherein performing segmentation further comprises:

determining a hypothesis about a candidate document;

displaying in a viewfinder the candidate document; and

checking the candidate document by applying at least one reliability criterion.

13. The method of claim 1, further comprising converting the frame into a grayscale frame, wherein the frame is a color frame.

14. The method of claim 13, wherein the frame is compressed using bilinear interpolation.

15. The method of claim 8, wherein the one or more search areas are rectangular areas of predefined dimensions.

16. The method of claim 8, wherein the one or more search areas are quarter circles with centers located in corners of the viewfinder with predefined radiuses.

17. The method of claim 8, wherein the smoothing of noise interferences is performed using at least one of:

a minimum filter, a maximum filter, and a medium filter.

18. The method of claim 8, wherein the differential edge detection method is operator based on differentials to pixels within regions of interest of the selected current view.

19. The method of claim 1, further comprising:

detecting corners for each candidate document within a current view; and

checking a reliability criteria related to corners for each candidate document.

20. The method of claim 19, wherein said detecting of corners is performed in a region of interest, wherein the region of interest is quarter circular area relative to a respective corner of the selected current view, one quarter circular in corners of a viewfinder.

21. The method of claim 1, wherein computing histograms further comprises:

detecting global extreme values in the histogram;

verifying detected global extreme values by at least one of the criteria;

generating cutting lines passing in a neighborhood of detected global extreme values;

segmenting analyzed current view into several areas along the cutting lines; and

analyzing at least one segmented area.

22. The method of claim 1, wherein analyzing at least one segmented area comprises applying information of gradient separators within each area, internal feature objects.

23. The method of claim 1, wherein selecting the frame in a viewfinder further comprises:

analyzing a detected document for the level of distortions; and

selecting current views with minimum level of distortions for the subsequent capturing.

24. The method of claim 8, wherein:

detecting edges includes detecting edges for each of a plurality of documents in the detected current view; and

checking the reliability criterion includes checking the reliability criterion for each of the plurality of documents in the detected current view.

25. A system comprising:

one or more data processors; and

one or more storage devices storing instructions that, when executed by the one or more data processors, cause the one or more data processors to perform operations comprising;

selecting a frame in a viewfinder;

converting the frame into a binary version of the frame;

computing a horizontal histogram of distribution of pixels along horizontal direction of the binarized version of the frame;

computing a vertical histogram of distribution of pixels along vertical direction of the binarized version of the frame;

performing segmentation of the binarized version of the frame using the horizontal histogram and the vertical histogram resulting in one or more horizontal and vertical segments, wherein each of the one or more horizontal and vertical segments corresponds to one document included in the frame;

identifying at least one document in the selected frame, wherein one or more parameters are associated with the at least one document;

comparing values of one or more parameters associated with the at least one identified document with predetermined values for the one or more parameters; and

capturing the at least one document upon determining that values of the one or more parameters associated with the document match predetermined values corresponding to the one or parameters.

26. The system of claim 25, wherein the frame selection occurs at a predetermined frequency.

27. The system of claim 26, wherein the predetermined frequency specifies a number of frames per second.

28. The system of claim 25, wherein the predetermined values of the one or more predetermined parameters are received from a user interface allowing a user to specify values for the predetermined parameters.

29. The system of claim 25, wherein the one or more predetermined parameters include geometrical proportions

30. The system of claim 25, wherein the one or more predetermined parameters include a parameter specifying a width to length ratio.

31. The system of claim 30, wherein the width to length ratio of the document matches the predetermined width to length ratio.

32. A computer-readable storage medium having machine instructions stored therein, the instructions being executable by a processor to cause the processor to perform operations comprising:

selecting a frame in a viewfinder;

converting the frame into a binary version of the frame;

computing a horizontal histogram of distribution of pixels along horizontal direction of the binarized version of the frame;

computing a vertical histogram of distribution of pixels along vertical direction of the binarized version of the frame;

performing segmentation of the binarized version of the frame using the horizontal histogram and the vertical histogram resulting in one or more horizontal and vertical segments, wherein each of the one or more horizontal and vertical segments corresponds to one document included in the frame;

identifying at least one document in the selected frame, wherein one or more parameters are associated with the at least one document;

comparing values of one or more parameters associated with the at least one identified document with predetermined values for the one or more parameters; and

capturing the at least one document upon determining that values of the one or more parameters associated with the document match predetermined values corresponding to the one or parameters.