Image analysis – Histogram processing – For setting a threshold
Reexamination Certificate
1999-03-05
2003-01-14
Wu, Jingge (Department: 2623)
Image analysis
Histogram processing
For setting a threshold
C382S176000, C382S112000, C358S464000
Reexamination Certificate
active
06507670
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to analyzing and improving video images, and specifically to a system and process for detecting discontinuities in text characters, such as those caused by dot matrix printing, and for improving the quality of such characters.
2. Description of the Prior Art
The ability to identify and locate relevant information within images of item-sized documents, such as financial documents (personal checks, business checks, etc.), is currently a feature of products offered by NCR Corporation, such as the NCR 7780, NCR ImageMark, and the NCR DP-ATM ImageMark. With respect to an image of a document such as a personal check, these products can identify and locate the various types of information contained therein, such as courtesy amount, legal amount, date, signature, payee, etc. The identified portions of the document can then be used in a variety of ways to help automate the processing of these documents. Of course, these systems can have a variety of applications, not just within the financial document processing field.
Several types of quality problems can occur in document images. For example, information can occur unexpectedly in an image which consequently does not match document or field models. Conversely, an image may be missing important expected information. These problems may occur for a number of reasons, including:
The image capture process may have failed.
The item may have been misaligned, skewed, or misoriented during image capture.
The image may have been poorly thresholded.
The image may contain excessive noise.
The item may contain additional unexpected printed information, such as an annotation on a personal check.
Preprinted item graphics may interfere with needed field information in the image.
There may be false or misleading information on the document.
With respect to the financial industry in particular, the ability to determine the quality of images of financial documents has become increasingly important as banks begin to migrate from a paper-based environment to an image-based one. Without a reliable measure of image quality, banks do not know if financial document images contain the same information as their paper counterparts. If information is missing from these images, such as a courtesy zone, required transactions involving this information may be negatively affected. By assessing the quality of a document, as well as the quality of fields on the document, the bank can determine the set of transactions for which the image can be used, and determine if the paper document must be retained.
There therefore exists a need in the art for an improved system that automatically and accurately evaluates the quality of a document image, rejects the document image if necessary, with feedback to the user, etc., and/or repairs the image so that the document can be properly handled.
SUMMARY OF THE INVENTION
A system is provided for removing a background pattern from a binary image. The system comprises a device for storing the binary image, and a processing device. The processing device may perform various steps, including (1) dividing, the binary image stored in the storage device into a plurality of stored image regions, (2) for each of the plurality of stored image regions, calculating the image density value associated with such stored image region, (3) counting the number of stored image regions associated with each image density value, (4) identifying the image density values having the maximum number of associated stored image regions, (5) selecting as a threshold density value a less dense value from the identified image density values, and (6) for each stored image region stored in the storage device having a density value less than the threshold density value, resetting the stored image region to a zero value.
REFERENCES:
patent: 5555556 (1996-09-01), Ozaki
patent: 5889885 (1999-03-01), Moed et al.
patent: 6018590 (2000-01-01), Gaborski
patent: 6081618 (2000-01-01), Naoi et al.
Wu et al. “Finding text in images” Proc. 2nd ACM International Conference on Digital Libraries, p. 3-12, Jul. 1997.*
Luijendijk “Automatic threshold selection using histograms based on the count of 4-connected regions” Pattern Recognition Letters 1991, vol. 12, No. 4, p. 219-28, Apr. 1991.
Needle & Rosenberg P.C.
Wu Jingge
LandOfFree
System and process for removing a background pattern from a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and process for removing a background pattern from a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and process for removing a background pattern from a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3037939