Numerical model for image feature extraction

• US 7,194,393 B2
• Filed: 01/19/2006
• Issued: 03/20/2007
• Est. Priority Date: 09/05/2001
• Status: Expired due to Term

• Current Assignee: Gemalto SA (Thales SA)
• Original Assignee: Cogent Systems, Inc.
• Inventors: Hsieh, Ming, Wei, Xiangshu
• Primary Examiner(s): PALADINI, ALBERT WILLIAM
• Application Number:
  US11/336,130
• Publication Number:
  US 20060115131A1
• Time in Patent Office: 425 Days
• Field of Search: 703/2, 382/190, 382/192, 382/195, 382/201, 382/206, 382/243, 382267-282, 382/286, 382/299, 382/300, 382/307
• US Class Current: 703/2
• CPC Class Codes: G06V 40/10 Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
• CPC Class Codes: G06V 40/1347 Preprocessing; Feature extraction

Abstract

In one embodiment, the invention is a method for generating geometric patterns from an image having a plurality of ridges and mesh points. The method includes dividing the image into a plurality of local regions with a surface appearance of fluid flow; establishing a mathematical model according to conditions of the plurality of local regions with a surface appearance of fluid flow; solving the mathematical model; and mapping respective values of the solutions of the mathematical model to respective local regions of the image.

Claims

1. A method performed by a computer for extracting features from an image, the method comprising:

dividing the image into a plurality of local regions with a surface appearance of fluid flow;

establishing a mathematical model according to conditions of the plurality of local regions with a surface appearance of fluid flow;

solving the mathematical model; and

mapping respective values of the solutions of the mathematical model to respective local regions of the image.

2. The method of claim 1, wherein the step of establishing a mathematical model comprises:

forming a partial differential equation;

determining initial conditions for the partial differential equation; and

determining boundary condition for the partial differential equation.

3. The method of claim 2, wherein the step of determining initial conditions comprises:

normalizing the image to reduce variations in gray-level values along ridges of the image;

estimating property values the image; and

mapping the estimated property values into weight coefficients of the partial differential equation.

4. The method of claim 2, wherein the step of forming a partial differential equation comprises:

calculating a plurality of intrinsic properties of the image according to image ridge pattern; and

mapping the plurality of intrinsic properties into coefficients of the partial differential equation.

5. The method of claim 2, wherein the step of determining initial conditions comprises:

normalizing the image to reduce variations in gray-level values along ridges of the image;

estimating property values of the image; and

mapping the estimated property values into weight coefficients of the partial differential equation.

6. The method of claim 2, wherein the step of determining boundary condition comprises:

drawing a close boundary within the image; and

setting boundary condition on the drawn boundary.

7. The method of claim 1, wherein the image is a fingerprint.

8. The method of claim 1, wherein the image is a facial image.

9. The method of claim 1, wherein the image is a hand-palm image.

10. The method of claim 1, wherein the image is an eye iris image.

11. The method of claim 1, wherein the image is a texture image.

12. The method of claim 1, wherein the image is an eye retina image.

13. A method performed by a computer for image processing of a biometric image, the method comprising:

dividing the biometric image into a plurality of local regions each with a flow-like pattern;

calculating an intrinsic property of each of the plurality of the local regions according to flow-like pattern to establish numerical relationship between visual appearance of each of the plurality of local regions; and

approximating solutions of a numerical model according to the established numerical relationship to determine features of the biometric image.

14. The method of claim 13, wherein the step of calculating an intrinsic property comprises:

forming a partial differential equation;

mapping the plurality of intrinsic properties into coefficients of the partial differential equation; and

determining a boundary condition for the partial differential equation from the image to establish a relationship between properties of the image and the partial differential equation.

15. The method of claim 14, wherein the step of forming a partial differential equation comprises:

determining initial conditions for the partial differential equation; and

determining boundary condition for the partial differential equation.

16. The method of claim 14, wherein the step of determining initial conditions comprises:

normalizing the image to reduce variations in gray-level values along ridges of the image;

estimating property values the image; and

mapping the estimated property values into weight coefficients of the partial differential equation.

17. The method of claim 14, wherein the step of determining boundary condition comprises:

drawing a close boundary within the image; and

setting boundary condition on the drawn boundary.

18. The method of claim 13, wherein the biometric image is one or more of a fingerprint image, a facial image, a hand-palm image, an eye iris image, a texture image, and an eye retina image.

19. A digital signal processor (DSP) having stored thereon a set of instructions including instruction for extracting features from an image and when executed by the DSP causing the DSP to perform the steps of:

establishing a mathematical model according to conditions of the plurality of local regions with a surface appearance of fluid flow;

solving the mathematical model; and

mapping respective values of the solutions of the mathematical model to respective local regions of the image.

20. The DSP of claim 19, wherein the image is one or more of a fingerprint image, a facial image, a hand-palm image, an eye iris image, a texture image, and an eye retina image.