Systems and methods for fingerprint matching based on transforming the deformable mesh

1. A fingerprint matching system comprising:

a memory; and

a processor operatively coupled to the memory, wherein the processor is configured to:

obtain first minutiae information of a first fingerprint;

obtain second minutiae information of a second fingerprint;

construct a deformable mesh based on the first minutiae;

transform a state of the deformable mesh, wherein the transforming is based at least in part on the second minutiae information of the second fingerprint;

transform the deformable mesh to minimize an energy and to generate a distortion-compensated first fingerprint, wherein the energy is defined to increase according to deformation of the deformable mesh, and decrease as corresponding minutiae approach one another, wherein the processor transforms the deformable mesh as follows:

analytically calculate first derivatives of the energy, wherein the first derivatives relate to a direction of transformation;

analytically calculate second derivatives of the energy, wherein the second derivatives relate to an amount of the transformation; and

transform the deformable mesh based on the first derivatives and the second derivatives; and

determine whether the first fingerprint matches the second fingerprint based on the distortion-compensated first fingerprint.

2. The fingerprint matching system of claim 1 , wherein the processor is further configured to warp the first minutiae according to the transformed state of the deformable mesh.

3. The fingerprint matching system of claim 1 , wherein the processor is further configured to determine correspondences between the first minutiae and the second minutiae, and wherein the transforming is based at least in part on the correspondences.

4. The fingerprint matching system of claim 1 , wherein the processor is further configured to determine locations of vertices in the deformable mesh so that the deformable mesh covers the first minutiae.

5. The fingerprint matching system of claim 4 , wherein the processor is further configured to: obtain edge lengths of edges connecting the vertices; and

determine the locations of the vertices based on the obtained edge lengths.

6. The fingerprint matching system of claim 4 , wherein the state of the deformable mesh is defined based on the locations of the vertices.

7. The fingerprint matching system of claim 1 , wherein the deformable mesh is a triangulated 2-dimensional mesh having hexagonally connected vertices.

8. The fingerprint matching system of claim 1 , wherein the first minutiae and the second minutiae include at least one of a ridge ending and a ridge bifurcation.

9. A system for compensating surface distortion of an input fingerprint comprising:

a memory; and

a processor operatively coupled to the memory and configured to:

obtain input minutiae information of an input fingerprint;

construct a deformable mesh based on the input minutiae information; and

transform a state of the deformable mesh to compensate surface distortion of the input fingerprint, wherein the transforming is based at least in part on template minutiae information associated with a template fingerprint;

transform the deformable mesh to minimize an energy, wherein the energy is defined to increase according to deformation of the deformable mesh, and decrease as corresponding minutiae approach one another, and wherein the processor is configured to transform the deformable mesh to minimize the energy as follows:

analytically calculate first derivatives of the energy, wherein the first derivatives relate to a direction of transformation;

analytically calculate second derivatives of the energy, wherein the second derivatives relate to an amount of the transformation; and

transform the deformable mesh based on the first derivatives and the second derivatives.

10. The system of claim 9 , wherein the processor is further configured to:

obtain the template minutiae information of the template fingerprint; and

determine correspondences between the input minutiae and the template minutiae, and

wherein the transforming of the state is based at least in part on the correspondences.

11. A fingerprint matching method comprising:

obtaining, by a computing device, minutiae of a first fingerprint and minutiae of a second fingerprint;

finding correspondences between the minutiae of the first fingerprint and the minutiae of the second fingerprint;

constructing a deformable mesh based on the minutiae of the first fingerprint;

transforming the deformable mesh to obtain a distortion-compensated first fingerprint based on the correspondences, wherein the transforming comprises transforming the deformable mesh to minimize an energy, and wherein the energy is defined to increase according to deformation of the deformable mesh, and decrease as corresponding minutiae approach one another, wherein transforming the deformable mesh to minimize the energy comprises:

analytically calculating first derivatives of the energy, wherein the first derivatives relate to a direction of transformation;

analytically calculating second derivatives of the energy, wherein the second derivatives relate to an amount of the transformation; and

transforming the deformable mesh based on the first derivatives and the second derivatives; and

determining whether the first fingerprint matches the second fingerprint based on the distortion-compensated first fingerprint.

12. The fingerprint matching method of claim 11 , wherein the energy disregards correspondences having a distance greater than a confidence radius, and further comprising:

attenuating the confidence radius;

determining whether the attenuated confidence radius satisfies a predetermined ending condition; and

if it is determined that the attenuated confidence radius does not satisfy the predetermined ending condition, repeating the operations of transforming the deformable mesh to minimize the energy, attenuating the confidence radius, and determining whether the attenuated confidence radius satisfies the predetermined ending condition.

13. The fingerprint matching method of claim 12 , wherein the predetermined ending condition is satisfied when the attenuated confidence radius is less than a predetermined threshold.

14. A fingerprint matching apparatus comprising:

means for obtaining minutiae of a first fingerprint and minutiae of a second fingerprint;

means for finding correspondences between the minutiae of the first fingerprint and the minutiae of the second fingerprint,

means for constructing a deformable mesh on the minutiae of the first fingerprint;

means for transforming the deformable mesh to obtain a distortion-compensated first fingerprint based on the correspondences, wherein the means for transforming comprises means for transforming the deformable mesh to minimize an energy, and wherein the energy is defined to increase according to deformation of the deformable mesh, and decrease as corresponding minutiae approach one another, wherein the means for transforming the deformable mesh to minimize the energy comprises means for analytically calculating first derivatives of the energy, means for analytically calculating second derivatives of the energy, and means for transforming the deformable mesh based on the first derivatives and the second derivatives, wherein the first derivatives relate to a direction of transformation, and wherein the second derivatives relate to an amount of the transformation; and

means for determining whether the first fingerprint matches the second fingerprint based on the distortion-compensated first fingerprint.

15. A method comprising:

obtaining, by a processor of a fingerprint matching system, first minutiae information of a first fingerprint;

obtaining, by the processor, second minutiae information of a second fingerprint;

constructing, by the processor, a deformable mesh based on the first minutiae;

transforming, by the processor, a state of the deformable mesh, wherein the transforming is based at least in part on the second minutiae information of the second fingerprint;

transforming, by the processor, the deformable mesh to minimize an energy and to generate a distortion-compensated first fingerprint, wherein the energy is defined to increase according to deformation of the deformable mesh, and decrease as corresponding minutiae approach one another, wherein the transforming comprises:

analytically calculating first derivatives of the energy, wherein the first derivatives relate to a direction of transformation;

analytically calculating second derivatives of the energy, wherein the second derivatives relate to an amount of the transformation; and

transforming the deformable mesh based on the first derivatives and the second derivatives; and

determining, by the processor, whether the first fingerprint matches the second fingerprint based on the distortion-compensated first fingerprint.