Automatic finger detection in document images

• US 7,634,112 B1
• Filed: 09/13/2004
• Issued: 12/15/2009
• Est. Priority Date: 09/13/2004
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Poncin, Guillaume Raphael, Khaliq, Siraj, Lefevere, Francois-Marie
• Primary Examiner(s): Bhatnagar; Anand
• Application Number:
  US10/938,505
• Time in Patent Office: 1,919 Days
• Field of Search: 382/173, 382/174, 382/176, 382/181, 382/190, 382/275, 382/100, 382/103, 382/112, 382/195, 382/274, 382/321, 358/448, 358/461, 358/452, 358/462, 358/500, 358/520
• US Class Current: 382/112
• CPC Class Codes: G06V 10/987 with the intervention of an operator
• CPC Class Codes: G06V 30/40 Document-oriented image-based pattern recognition
• CPC Class Codes: G06V 40/107 Static hand or arm
• CPC Class Codes: H04N 1/00488 providing an audible output to the user
• CPC Class Codes: H04N 1/0049 providing a visual indication to the user, e.g. using a lamp H04N1/00408 takes precedence
• CPC Class Codes: H04N 1/00795 Reading arrangements details of scanning heads H04N1/024; scanning arrangements therefor H04N1/04
• CPC Class Codes: H04N 1/00801 according to characteristics of the original
• 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 2201/0098 User intervention not otherwise provided for, e.g. placing documents, responding to an alarm
• CPC Class Codes: H04N 2201/0434 specially adapted for scanning pages of a book

Abstract

Techniques for scanning a document, such as a book, magazine, or catalog, are described in which artifacts (e.g., a hand or other body part) in the scanned image are automatically detected. When the artifact is determined to be in a critical portion of the image, a signal may be issued to re-scan the image.

Claims

1. A method performed by a processor associated with one or more cameras, the method comprising:

locating, using the processor, an image portion that represents a human body part in an image representing a portion of a page of a document;

expanding, using the processor, the located image portion that represents a human body part to include additional image portions representing the human body part;

determining, using the processor, whether the expanded image portion is in a particular portion of the image; and

issuing, using the processor, a signal to re-scan the page when the located image portion that represents a human body part is determined to be in a particular portion of the image.

2. The method of claim 1, where the determining whether the expanded image portion that represents a human body part is in a particular portion of the image is based on whether the expanded image portion is in a portion of the image needed for later processing of the image.

3. The method of claim 1, where the document is a book, magazine, newspaper, catalog, or poster.

4. The method of claim 1, where the signal is an audible or visual signal for an operator.

5. The method of claim 1, where the human body part is a hand or a portion of a hand or wrist.

6. The method of claim 1, where the determination of whether the expanded image portion that represents a human body part is in a particular portion of the image includes:

determining whether the expanded image portion occludes a text area of the image.

7. The method of claim 1, where the determination of whether the expanded image portion that represents a human body part is in a particular portion of the image includes:

determining whether the expanded image portion will obstruct later optical character recognition (OCR) processing of the image.

8. The method of claim 1, where locating the image portion that represents a human body part in the image includes:

locating image portions that represent areas of skin in the image based on hue and saturation values of pixels of the image.

9. The method of claim 8, where a range of hue and saturation values are selected to correspond to skin.

10. The method of claim 1, where determining whether the expanded image portion that represents a human body part is in a particular portion of the image includes:

distinguishing between text and non-text areas of the image.

11. The method of claim 10, where determining whether the expanded image portion that represents a located human body part is in a particular portion of the image further includes:

constructing a histogram of value attributes from a hue, saturation, value (HSV) color model; and

locating the text and non-text areas of the image as areas corresponding to portions of the histogram.

12. A system comprising:

a camera to generate an image of a page of a document;

a memory comprising instructions to;

locate an image portion that represents a human body part in the image of the page,

expand the located image portion that represents a human body part to include additional image portions representing the human body part,

determine whether the expanded image portion that represents a human body part is in a particular portion of the image, and

issue a signal to the camera to re-scan the page when the expanded image portion that represents a human body part is determined to be in a particular portion of the image; and

a processor coupled to the memory, to execute the instructions.

13. The system of claim 12, further comprising:

a stereoscopic camera to generate images of the page of the document from which a three-dimensional image of a surface of the page of the document is generated.

14. The system of claim 12, where the document is a book, magazine, newspaper, catalog, or poster.

15. The system of claim 12, where the instructions to locate the human body part include instructions to locate the human body part based on a selected range of hue and saturation values associated with the human body part.

16. The system of claim 12, where the signal is an audible or visual signal for an operator.

17. The system of claim 12, where the human body part is a hand or a portion of a hand or wrist.

18. The system of claim 12, where the memory further comprises instructions to:

determine whether the expanded image portion that represents a human body part occludes a text area of the image.

19. The system of claim 12, where the memory further comprises instructions to:

determine whether the expanded image portion that represents a human body part will obstruct later optical character recognition (OCR) processing of the image.

20. The system of claim 12, where the memory further comprises instructions to:

locate image portions that represent areas of skin in the image based on hue and saturation values of pixels of the image.

21. The system of claim 12, where the memory further comprises instructions to:

distinguish between text and non-text areas of the image.

22. The system of claim 21 where the memory further comprises instructions to:

construct a histogram of value attributes from a hue, saturation, value (HSV) color model; and

locate the text and non-text areas of the image as areas corresponding to portions of the histogram.

23. A device comprising:

means for locating an image portion that represents a human body part in an image representing at least a portion of a page of a document;

means for expanding the located image portion that represents a human body part to include additional image portions representing the human body part;

means for determining whether the expanded image portion that represents a human body part is in a text portion of the image; and

means for issuing a signal to re-scan the page when the located human body part is determined to not be in a text portion of the image.

24. The device of claim 23, where the means for determining whether the expanded image portion represents a human body part comprises:

means for constructing a histogram of value attributes from a hue, saturation, value (HSV) color model; and

means for locating text and non-text areas of the image as areas corresponding to portions of the histogram.

25. The device of claim 23, where the means for determining whether the expanded image portion represents a human body part comprises:

means for determining whether the expanded image portion will obstruct later optical character recognition (OCR) processing of the image.