Fingerprint verification method and fingerprint verification device

• US 7,349,559 B2
• Filed: 12/18/2003
• Issued: 03/25/2008
• Est. Priority Date: 12/27/2002
• Status: Active Grant

• Current Assignee: 138 East LCD Advancements Limited (Vector Capital Corporation)
• Original Assignee: Seiko Epson Corporation (Seiko Group)
• Inventors: Miyasaka, Mitsutoshi
• Primary Examiner(s): Wu; Jingge
• Assistant Examiner(s): Lee; John W
• Application Number:
  US10/738,050
• Publication Number:
  US 20040184642A1
• Time in Patent Office: 1,559 Days
• Field of Search: 382124-127
• US Class Current: 382/124
• CPC Class Codes: G06V 40/1365 Matching; Classification

Abstract

A fingerprint verification method includes extracting N_D sample feature points from a fingerprint image that has been sampled, assigning sample numerical values to each of the sample feature points as sample numerical value information. The sample numerical values are obtained from a plurality of sample connection lines, which connect each of the sample feature points with L_D close sample feature points, which are close to each of the sample feature points. Based on results of comparison between the sample numerical value information and registration numerical value information, which is registration numerical values assigned to each of N_R registration feature points, associating m sample feature points among the N_D sample feature points with m registration feature points among the N_R registration feature points. A position relationship among the m sample feature points is compared with a position relationship among the m registration feature points.

Claims

1. A fingerprint verification method using a fingerprint image, the method comprising:

extracting N_D sample feature points (N_D is an integer greater than or equal to four) from a fingerprint image that has been sampled;

assigning sample numerical values to each of the sample feature points as sample numerical value information, the sample numerical values being obtained from a plurality of sample connection lines, which connect each of the sample feature points with L_D close sample feature points (2≦

L_D≦

N_D−

1, L_D is an integer) which are close to each of the sample feature points;

comparing each of the sample numerical values assigned to each of the sample feature points with each of registration numerical values assigned to each of N_R registration feature points (N_R is an integer greater than or equal to four), and assigning a registration feature point among the N_R registration feature points which corresponds to a registration numerical value that coincides with a sample numerical value among the sample numerical values to the coincided sample numerical value;

deleting a sample feature point among the N_D sample feature points which has sample numerical values that have not been assigned with the registration feature points at a percentage equal to or greater than a first percentage;

associating one registration feature point among the assigned registration feature points with only a sample feature point among the remaining sample feature points excluding the deleted sample feature point, the sample feature point including sample numerical values that have been assigned in common at a second percentage or more to the one register feature point; and

when the number of sample feature points associated with the registration feature points is m or more (m is an integer defined as 3≦

m≦

N_D and 3≦

m≦

N_R), comparing a position relationship among the m or more sample feature points with the position relationship among the registration feature points associated with the m or more sample feature points.

2. The fingerprint verification method as defined in claim 1, comprising:

assigning the registration numerical values to each of the registration feature points, the registration numerical values being obtained from a plurality of registration connection lines, which connect each of the N_R registration feature points extracted from a fingerprint image to be registered with L_R close registration feature points (2≦

L_R≦

N_R−

1, L_R is an integer) which are close to each of the registration feature points.

3. The fingerprint verification method as defined in claim 2, wherein L_R is equal to or greater than L_D.

4. The fingerprint verification method as defined in claim 2,

wherein registration numerical values include lengths of the registration connection lines, and

wherein the sample numerical values include lengths of the sample connection lines.

5. The fingerprint verification method as defined in claim 2,

wherein the registration numerical values include angles formed by the registration connection lines, and

wherein the sample numerical values include angles formed by the sample connection lines.

6. The fingerprint verification method as defined in claim 2,

wherein the registration numerical values include lengths of the registration connection lines and angles formed by the registration connection lines, and

wherein the sample numerical values include lengths of the sample connection lines and angles formed by the sample connection lines.

7. The fingerprint verification method as defined in claim 2,

wherein, in a case where the registration connection lines connected at each of the registration feature points as a starting point are registration vectors starting from each of the registration feature points, the registration numerical values include inner products formed by the registration vectors, and

wherein, in a case where the sample connection lines connected at each of the sample feature points as a starting point are sample vectors starting from each of the sample feature points, the sample numerical values include inner products formed by the sample vectors.

8. The fingerprint verification method as defined in claim 2,

wherein the registration numerical values include numbers of fingerprint ridges which intersect the registration connection lines, and

wherein the sample numerical values include numbers of fingerprint ridges which intersect the sample connection lines.

9. The fingerprint verification method as defined in claim 2,

wherein the registration numerical values include numbers of fingerprint ridges which intersect the registration connection lines and angles formed by the registration connection lines, and

wherein the sample numerical values include numbers of fingerprint ridges which intersect the sample connection lines and angles formed by the sample connection lines.

10. The fingerprint verification method as defined in claim 2,

wherein, in a case where registration ridge vectors starting from each of the registration feature points are defined by using numbers of fingerprint ridges which intersect the registration connection lines as vector lengths, the registration numerical values include inner products formed by the registration ridge vectors, and

wherein, in a case where sample ridge vectors starting from each of the sample feature points are defined by using numbers of fingerprint ridges which intersect the sample connection lines as vector lengths, the sample numerical values include inner products formed by the sample ridge vectors.

11. The fingerprint verification method as defined in claim 2,

wherein a real number unit value is assigned to one of the registration feature points in a case where the one registration feature point is a first feature point, and an imaginary number unit value is assigned to the one registration feature point in a case where the one registration feature point is a second feature point,

wherein the registration numerical values include a product of the real number unit value or the imaginary number unit value assigned to each of the registration feature points located on both ends of each of the registration connection lines and a length of each of the registration connection lines,

wherein a real number unit value is assigned to one of the sample feature points in a case where the one sample feature point is the first feature point, and an imaginary number unit value is assigned to the one sample feature point in a case where the one sample feature point is the second feature point, and

wherein the sample numerical values include a product of the real number unit value or the imaginary number unit value assigned to each of the sample feature points located on both ends of each of the sample connection lines and a length of each of the sample connection lines.

12. The fingerprint verification method as defined in claim 2,

wherein a real number unit value is assigned to one of the registration feature points in a case where the one registration feature point is a first feature point, and an imaginary number unit value is assigned to the one registration feature point in a case where the one registration feature point is a second feature point,

wherein the registration numerical values include a product of an angle formed by two of the registration connection lines and the real number unit value or the imaginary number unit value assigned to each of the registration feature points located on both ends of the two registration connection lines,

wherein a real number unit value is assigned to one of the sample feature points in a case where the one sample feature point is the first feature point, and an imaginary number unit value is assigned to the one sample feature point in a case where the one sample feature point is the second feature point, and

wherein the sample numerical values include a product of an angle formed by two of the sample connection lines and the real number unit value or the imaginary number unit value assigned to each of the sample feature points located on both ends of the two sample connection lines.

13. The fingerprint verification method as defined in claim 2,

wherein a real number unit value is assigned to one of the registration feature points in a case where the one registration feature point is a first feature point, and an imaginary number unit value is assigned to the one registration feature point in a case where the one registration feature point is a second feature point,

wherein the registration numerical values include a product of the real number unit value or the imaginary number unit value assigned to each of the registration feature points located on both ends of each of the registration connection lines and a length of each of the registration connection lines and a product of an angle formed by two of the registration connection lines and the real number unit value or the imaginary number unit value assigned to each of the registration feature points located on both ends of the two registration connection lines,

wherein a real number unit value is assigned to one of the sample feature points in a case where the one sample feature point is the first feature point, and an imaginary number unit value is assigned to the one sample feature point in a case where the one sample feature point is the second feature point, and

wherein the sample numerical values include a product of the real number unit value or the imaginary number unit value assigned to each of the sample feature points located on both ends of each of the sample connection lines and a length of each of the sample connection lines and a product of an angle formed by two of the sample connection lines and the real number unit value or the imaginary number unit value assigned to each of the sample feature points located on both ends of the two sample connection lines.

14. The fingerprint verification method as defined in claim 2,

wherein a real number unit value is assigned to one of the registration feature points in a case where the one registration feature point is a first feature point, and an imaginary number unit value is assigned to the one registration feature point in a case where the one registration feature point is a second feature point,

wherein, in a case where the registration connection lines connected at each of the registration feature points as a starting point are registration vectors starting from each of the registration feature points, the registration numerical values include a product of inner products formed by the registration vectors and the real number unit value or the imaginary number unit value assigned to each of the registration feature points located at a starting point and an end point of the registration vectors,

wherein a real number unit value is assigned to one of the sample feature points in a case where the one sample feature point is the first feature point, and an imaginary number unit value is assigned to the one sample feature point in a case where the one sample feature point is the second feature point, and

wherein, in a case where the sample connection lines connected at each of the sample feature points as a starting point are sample vectors starting from each of the sample feature points, the sample numerical values include a product of inner products formed by the sample vectors and the real number unit value or the imaginary number unit value assigned to each of the sample feature points located at a starting point and an end point of the sample vectors.

15. The fingerprint verification method as defined in claim 2,

wherein a real number unit value is assigned to one of the registration feature points in a case where the one registration feature point is a first feature point, and an imaginary number unit value is assigned to the one registration feature point in a case where the one registration feature point is a second feature point,

wherein the registration numerical values include a product of the real number unit value or the imaginary number unit value assigned to each of the registration feature points located on both ends of each of the registration connection lines and number of fingerprint ridges which intersect the registration connection lines,

wherein a real number unit value is assigned to one of the sample feature points in a case where the one sample feature point is the first feature point, and an imaginary number unit value is assigned to the one sample feature point in a case where the one sample feature point is the second feature point, and

wherein the sample numerical values include a product of the real number unit value or the imaginary number unit value assigned to each of the sample feature points located on both ends of each of the sample connection lines and number of fingerprint ridges which intersect the sample connection lines.

16. The fingerprint verification method as defined in claim 2,

wherein a real number unit value is assigned to one of the registration feature points in a case where the one registration feature point is a first feature point, and an imaginary number unit value is assigned to the one registration feature point in a case where the one registration feature point is a second feature point,

wherein, in a case where registration ridge vectors starting from each of the registration feature points are defined by using numbers of fingerprint ridges which intersect the registration connection lines as vector lengths, the registration numerical values include a product of inner products formed by the registration ridge vectors and the real number unit value or the imaginary number unit value assigned to each of the registration feature points located at a starting point and an end point of each of the registration vectors,

wherein a real number unit value is assigned to one of the sample feature points in a case where the one sample feature point is the first feature point, and an imaginary number unit value is assigned to the one sample feature point in a case where the one sample feature point is the second feature point, and

wherein, in a case where sample ridge vectors starting from each of the sample feature points are defined by using numbers of fingerprint ridges which intersect the sample connection lines as vector lengths, the sample numerical values include a product of inner products formed by the sample ridge vectors and the real number unit value or the imaginary number unit value assigned to each of the sample feature points located at a starting point and an end point of each of the sample vectors.

17. The fingerprint verification method as defined in claim 11, wherein the first feature point is a terminal point of a fingerprint ridge, and the second feature point is a branch point of a fingerprint ridge.

18. The fingerprint verification method as defined in claim 11, wherein the first feature point is a branch point of a fingerprint ridge, and the second feature point is a terminal point of a fingerprint ridge.