Living body information registration method, biometrics authentication method, and biometrics authentication apparatus

• US 8,660,318 B2
• Filed: 04/17/2012
• Issued: 02/25/2014
• Est. Priority Date: 10/30/2009
• Status: Active Grant

• Current Assignee: Fujitsu Frontech Limited (Fujitsu Limited)
• Original Assignee: Fujitsu Frontech Limited (Fujitsu Limited)
• Inventors: Komura, Kazuhiro, Takamatsu, Hiroyuki
• Primary Examiner(s): Mehta, Bhavesh
• Assistant Examiner(s): Harandi, Siamak
• Application Number:
  US13/448,872
• Publication Number:
  US 20120201431A1
• Time in Patent Office: 679 Days
• Field of Search: None
• US Class Current: 382/115
• CPC Class Codes: G06V 10/75 Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
• CPC Class Codes: G06V 40/12 Fingerprints or palmprints
• CPC Class Codes: G06V 40/14 Vascular patterns
• CPC Class Codes: G06V 40/67 by interactive indications to the user

Abstract

A vein authentication apparatus registers vein data and vein characteristic amounts, determines comparison order at vein data identification time by sorting records in descending order of degree of similarity on the basis of vein characteristic amounts for comparison and the registered vein characteristic amounts, and makes vein data comparison in accordance with the comparison order. The vein authentication apparatus uses 32 frequencies obtained by performing a Fourier transform, 12 angles obtained by performing a Fourier transform, curvature directions in 36 angle areas, segment directions in 18 angle areas, and a vein amount as the vein characteristic amounts. By doing so, the accuracy of comparison order is improved and a comparison using a record for which a degree of similarity is low can be omitted. As a result, the speed of a comparison process and the maximum number of pieces of living body information which can be registered can be increased.

Claims

1. A living body information registration method executed by a computer, the method comprising:

extracting vein data and characteristic amounts from a captured image which is living body information; and

registering the extracted vein data and the extracted characteristic amounts,

the characteristic amounts including;

a first vein characteristic amount which is a frequency component obtained by characterizing periodicity of veins by performing a Fourier transform on a vein image obtained by reducing the vein data;

a second vein characteristic amount which is an angle component obtained by characterizing a direction of a vein by performing a Fourier transform on the vein image;

a third vein characteristic amount which is a curvature direction component obtained by characterizing a direction of a curvature of a vein of the vein image;

a fourth vein characteristic amount which is a segment direction component obtained by characterizing a direction of a vein segment obtained by dividing a vein of the vein image; and

a fifth vein characteristic amount which is a vein amount obtained by characterizing an amount of veins included in the vein image,

wherein the fourth vein characteristic amount indicates distribution of segment direction components obtained by;

dividing the vein image into a plurality of areas in accordance with a first division pattern;

for each area, vectorizing the vein image, finding a vein segment made up of straight segments bent at vector points, calculating a slope of each segment, and drawing a histogram;

dividing the vein image into a plurality of areas in accordance with a second division pattern that specifies different division positions from the first division pattern; and

for each area obtained under the second division pattern, vectorizing the vein image, finding a vein segment made up of straight segments bent at vector points, calculating a slope of each segment, and drawing a histogram, and

wherein the fifth vein characteristic amount indicates a distribution of vein amounts obtained by;

dividing the vein image into a plurality of areas in accordance with the first division pattern;

for each area, calculating the number of pixels corresponding to the veins included in the vein image and drawing a histogram;

dividing the vein image into a plurality of areas in accordance with the second division pattern; and

for each area obtained under the second division pattern, calculating the number of pixels corresponding to the veins included in the vein image and drawing a histogram.

2. A biometrics authentication method executed by a computer, the method comprising:

extracting vein data and characteristic amounts from a captured image which is living body information;

acquiring registered vein data and registered characteristic amounts in each record registered in advance;

calculating scores indicative of rough degrees of similarity from the extracted characteristic amounts and the acquired registered characteristic amounts;

determining comparison order from a result of sorting the scores; and

comparing the vein data with the registered vein data in accordance with the determined comparison order,

the characteristic amounts and the registered characteristic amounts including;

a first vein characteristic amount which is a frequency component obtained by characterizing periodicity of veins by performing a Fourier transform on a vein image obtained by reducing the vein data;

a second vein characteristic amount which is an angle component obtained by characterizing a direction of a vein by performing a Fourier transform on the vein image;

a third vein characteristic amount which is a curvature direction component obtained by characterizing a direction of a curvature of a vein of the vein image;

a fourth vein characteristic amount which is a segment direction component obtained by characterizing a direction of a vein segment obtained by dividing a vein of the vein image; and

a fifth vein characteristic amount which is a vein amount obtained by characterizing an amount of veins included in the vein image,

wherein the fourth vein characteristic amount indicates distribution of segment direction components obtained by;

dividing the vein image into a plurality of areas in accordance with a first division pattern;

for each area, vectorizing the vein image, finding a vein segment made up of straight segments bent at vector points, calculating a slope of each segment, and drawing a histogram;

dividing the vein image into a plurality of areas in accordance with a second division pattern that specifies different division positions from the first division pattern; and

for each area obtained under the second division pattern, vectorizing the vein image, finding a vein segment made up of straight segments bent at vector points, calculating a slope of each segment, and drawing a histogram, and

wherein the fifth vein characteristic amount indicates a distribution of vein amounts obtained by;

dividing the vein image into a plurality of areas in accordance with the first division pattern;

for each area, calculating the number of pixels corresponding to the veins included in the vein image and drawing a histogram;

dividing the vein image into a plurality of areas in accordance with the second division pattern; and

for each area obtained under the second division pattern, calculating the number of pixels corresponding to the veins included in the vein image and drawing a histogram.

3. The biometrics authentication method according to claim 2, wherein the scores are total scores obtained by calculating a sum total of differences between the extracted characteristic amounts and the acquired registered characteristic amounts for each of the first through fifth vein characteristic amounts and calculating a sum total of the sum totals for the first through fifth vein characteristic amounts.

4. The biometrics authentication method according to claim 3, wherein the comparison order is determined by sorting the records registered in advance by the total scores.

5. The biometrics authentication method according to claim 2, wherein the vein data is compared with a part of the registered vein data which ranks high in the comparison order determined from the result of sorting the scores.

6. The biometrics authentication method according to claim 2, wherein records for which degrees of similarity are greater than or equal to a determined value are sorted by the degrees of similarity and the determination that a record which ranks highest is a record for a user whose veins are image-captured is made.

7. The biometrics authentication method according to claim 2, wherein the first vein characteristic amount is a frequency component indicative of the periodicity of the veins which is obtained by finding a spatial frequency component by performing a two-dimensional fast Fourier transform on the vein image, calculating a power spectrum from the spatial frequency component, expressing the power spectrum in polar coordinates, and finding energy in a doughnut-shaped area a center of which is an origin.

8. The biometrics authentication method according to claim 2, wherein the second vein characteristic amount is a direction component which indicates the direction in which the vein extends and which is obtained by finding a spatial frequency component by performing a two-dimensional fast Fourier transform on the vein image, calculating a power spectrum from the spatial frequency component, expressing the power spectrum in polar coordinates, and finding energy in a determined angle range.

9. The biometrics authentication method according to claim 2, wherein the third vein characteristic amount is a segment curvature direction component obtained by vectorizing the vein image, finding a vein segment made up of straight segments bent at vector points, calculating a curvature direction for two adjacent segments both ends of which are vector points that are a certain distance away from each other, and drawing a histogram.

10. The biometrics authentication method according to claim 9, wherein:

the vein image is divided into a plurality of areas in accordance with two division patterns which differ in division position; and

the histogram is drawn for each area.

11. A biometrics authentication apparatus for authenticating a user by living body information, the apparatus comprising:

an imaging device configured to capture an image of the user'"'"'s veins;

a vein database configured to store vein data and characteristic amounts included in the image which is the living body information and which is captured by the imaging device as registered vein data and registered characteristic amounts; and

one or more processors configured to perform a procedure comprising;

extracting the vein data and the characteristic amounts from the image, the characteristic amounts including;

a first vein characteristic amount which is a frequency component obtained by characterizing periodicity of veins by performing a Fourier transform on a vein image obtained by reducing the vein data;

a second vein characteristic amount which is an angle component obtained by characterizing a direction of a vein by performing a Fourier transform on the vein image;

a third vein characteristic amount which is a curvature direction component obtained by characterizing a direction of a curvature of a vein of the vein image;

a fourth vein characteristic amount which is a segment direction component obtained by characterizing a direction of a vein segment obtained by dividing a vein of the vein image; and

a fifth vein characteristic amount which is a vein amount obtained by characterizing an amount of veins included in the vein image;

registering the extracted vein data and the extracted characteristic amounts in the vein database as the registered vein data and the registered characteristic amounts;

acquiring the registered vein data and the registered characteristic amounts from the vein database;

calculating scores indicative of rough degrees of similarity from the extracted characteristic amounts and the registered characteristic amounts acquired, and determining comparison order from a result of sorting the scores; and

comparing the extracted vein data with the registered vein data acquired in accordance with the determined comparison order,

wherein the fourth vein characteristic amount indicates distribution of segment direction components obtained by;

dividing the vein image into a plurality of areas in accordance with a first division pattern;

for each area, vectorizing the vein image, finding a vein segment made up of straight segments bent at vector points, calculating a slope of each segment, and drawing a histogram;

dividing the vein image into a plurality of areas in accordance with a second division pattern that specifies different division positions from the first division pattern and

for each area obtained under the second division pattern, vectorizing the vein image, finding a vein segment made up of straight segments bent at vector points, calculating a slope of each segment, and drawing a histogram, and

wherein the fifth vein characteristic amount indicates a distribution of vein amounts obtained by;

dividing the vein image into a plurality of areas in accordance with the first division pattern;

for each area, calculating the number of pixels corresponding to the veins included in the vein image and drawing a histogram;

dividing the vein image into a plurality of areas in accordance with the second division pattern; and

for each area obtained under the second division pattern, calculating the number of pixels corresponding to the veins included in the vein image and drawing a histogram.