Generation of quality field information in the context of image processing

• US 7,155,040 B2
• Filed: 06/29/2004
• Issued: 12/26/2006
• Est. Priority Date: 06/29/2004
• Status: Active Grant

• Current Assignee: Bio-Key International Incorporated
• Original Assignee: Bio-Key International Incorporated
• Inventors: Nikiforov, Andrey
• Primary Examiner(s): Mancuso, Joseph
• Assistant Examiner(s): LIEW, ALEX KOK SOON
• Application Number:
  US10/879,603
• Publication Number:
  US 20050286801A1
• Time in Patent Office: 910 Days
• Field of Search: 382/125
• US Class Current: 382/125
• CPC Class Codes: G06V 10/993 Evaluation of the quality of the acquired pattern
• CPC Class Codes: G06V 40/1347 Preprocessing; Feature extraction

Abstract

A computer-implemented method is provided for generating a quality field representation of a collection of image data. The method includes receiving the collection of image data and identifying a target point along a ridge in the collection of image data. A directional field representation is estimated by assigning an angle value to the ridge indicative of an angle with respect to a reference line. Also, a covariance between the target point and a corresponding point along a line orthogonal to the angle value is calculated.

Claims

1. A computer-implemented method for generating a quality field representation of a collection of image data, the method comprising:

receiving the collection of image data;

identifying a target point along a ridge in the collection of image data;

estimating a directional field representation by assigning an angle value to the ridge indicative of an angle with respect to a reference line;

calculating a covariance between the target point and a corresponding point along a line orthogonal to the angle value; and

calculating quality values of a plurality of points proximate the target point and calculating the quality field representation based on the quality values.

2. The method of claim 1 and further comprising:

calculating an average among the quality values and the quality field representation and assigning the quality field representation based on the average.

3. A computer-implemented method for generating a quality field representation of a collection of image data, the method comprising:

receiving the collection of image data;

identifying a target point along a ridge in the collection of image data;

estimating a directional field representation by assigning an angle value to the ridge indicative of an angle with respect to a reference line;

calculating a covariance between the target point and a corresponding point along a line orthogonal to the angle value; and

calculating at least one of a period field representation and a directional field representation of the target point using the quality field representation.

4. A method of processing image data, comprising:

identifying a plurality of minutiae points within the image data; and

assigning quality values to each of the plurality of minutiae points, wherein each quality value is assigned by calculating a covariance between each respective minutiae point and a corresponding point displaced from each respective minutiae point and determining whether each covariance is indicative of a quasi-periodic structure.

5. A method of processing image data, comprising:

identifying a plurality of minutiae points within the image data;

assigning quality values to each of the plurality of minutiae points, wherein each quality value is assigned by calculating a covariance between each respective minutiae point and a corresponding point displaced from each respective minutiae point; and

calculating at least one of periodic values and direction values for each of the plurality of minutiae points based on the quality values.

6. The method of claim 5 and further comprising:

generating a model indicative of the image and including the quality values and the at least one of periodic values and direction values for the plurality of minutiae points.

7. The method of claim 6 and further comprising:

comparing the model with another model to determine a similarity measure.