System and process for assessing the quality of a signature within a binary image

• US 6,363,162 B1
• Filed: 03/05/1999
• Issued: 03/26/2002
• Est. Priority Date: 03/05/1998
• Status: Expired due to Term

• Current Assignee: NCR Atleos Corp.
• Original Assignee: NCR Corporation (NCR Voyix Corporation)
• Inventors: Moed, Michael E., Mkaouar, Mohamed
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): CHOOBIN, BARRY
• Application Number:
  US09/263,682
• Time in Patent Office: 1,117 Days
• Field of Search: 382/112, 382/187, 382/113, 382/118, 382/100
• US Class Current: 382/112
• CPC Class Codes: G06T 2207/30168 Image quality inspection
• CPC Class Codes: G06T 2207/30176 Document
• CPC Class Codes: G06T 5/30 Erosion or dilatation, e.g. thinning
• CPC Class Codes: G06T 5/77 Retouching; Inpainting; Scratch removal
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/133 Evaluation of quality of the acquired characters
• CPC Class Codes: G06V 30/164 Noise filtering
• CPC Class Codes: G06V 40/30 Writer recognition; Reading and verifying signatures

Abstract

A system and process is provided for assessing the quality of a signature within a binary image. The system comprises a mechanism for storing the binary image and a processor for performing various steps with respect to the image, including (1) detecting a background pattern within the binary image, (2) if a background pattern is detected, removing the background image from the binary image, thereby creating a cleaned binary image, (3) detecting discontinuities within a signature portion of the cleaned binary image, and (4) generating a measure of the signature quality for the binary image.

Claims

1. A system for assessing the quality of a signature within a binary image, comprising:

(a) means for storing the binary image;

(b) processing means for performing the steps of;

(i) detecting a background pattern within the binary image stored in the storing means;

(ii) if a background pattern is detected in step (i), removing the background image from the binary image, thereby creating a cleaned binary image;

(iii) detecting discontinuities within a signature portion of the cleaned binary image;

(iv) responsive to step (iii), generating a measure of the signature quality for the binary image.

2. The system of claim 1, wherein the processing means removes the background image by classifying the background pattern and applying an appropriate method for removing the background image, depending on the classification.

3. The system of claim 2, wherein the processing means classifies the background image based on a number of local maxima in a derivative of a horizontal projection of the binary image.

4. The system of claim 2, wherein the processing means uses density filtering, morphological filtering, or random pattern removal to remove the background image, depending on the classification of the background pattern.

5. The system of claim 4, wherein the processing means uses density filtering to remove a background image containing patterns that are lighter than the signature.

6. The system of claim 4, wherein the processing means uses morphological filtering to remove a background image containing linear patterns.

7. The system of claim 4, wherein the processing means uses random pattern removal to remove a background image containing short, random curves.

8. A method for assessing the quality of a signature within a binary image, comprising the steps of:

(a) storing the binary image;

(b) processing the binary image, including performing the steps of;

(i) detecting a background pattern within the stored binary image;

(ii) if a background pattern is detected in step (i), removing the background image from the binary image, thereby creating a cleaned binary image;

(iii) detecting discontinuities within a signature portion of the cleaned binary image;

(iv) responsive to step (iii), generating a measure of the signature quality for the binary image.

9. The method of claim 8, wherein the step of removing the background image includes classifying the background pattern and applying an appropriate method for removing the background image, depending on the classification.

10. The method of claim 9, wherein the step of classifying includes determining the number of local maxima in a derivative of a horizontal projection of the binary image.

11. The method of claim 9, wherein the step of removing the background image uses density filtering, morphological filtering, or random pattern removal to remove the background image, depending on the classification of the background pattern.

12. The method of claim 11, wherein density filtering is used to remove a background image containing patterns that are lighter than the signature.

13. The method of claim 11, wherein morphological filtering is used to remove a background image containing linear patterns.

14. The method of claim 11, wherein random pattern removal is used to remove a background image containing short, random curves.