IRIS RECOGNITION SYSTEM, A METHOD THEREOF, AND AN ENCRYPTION SYSTEM USING THE SAME

• US 20080219515A1
• Filed: 03/09/2007
• Published: 09/11/2008
• Est. Priority Date: 03/09/2007
• Status: Active Grant

• Current Assignee: Hongbog Corp.
• Original Assignee: Jiris USA Incorporated
• Inventors: Namgoong, Jong
• Granted Patent: US 8,023,699 B2
• Time in Patent Office: Days
• US Class Current: 382/117
• CPC Class Codes: G06V 40/193 Preprocessing; Feature extraction

Abstract

The present invention generally relates to an iris recognition system, a method thereof, and an encryption system, and more specifically, to an iris recognition system comprising: an image selector for selecting one image from a plurality of iris images inputted from an image input device; an internal edge extractor for extracting an internal edge of the selected iris image; an external edge extractor for extracting an external edge of the selected iris image; an iris area normalizer for normalizing an iris area; a security level applier for securing the iris which increases according to age; a rotational image processor for taking charge of processing for an inclined iris image; a feature value extractor for extracting a feature value from the iris image; and a consistency decider for comparing the extracted feature value with a stored iris code to decide whether the compared feature value and iris code are consistent; and wherein with regards to each of inputted iris images, the image selector scans each iris image in pixel unit by using a mask defined into a second area which is in square shape and a first area configured as the peripheral girth of the second area, calculates the number of pixels C1 that luminance values of pixels located in the first area are smaller than a first threshold value and the number of pixels C2 that luminance values of pixels located in the second area are bigger than a second threshold value, and selects an image of which the calculated pixel C2 values are minimum.

Claims

1. An iris recognition system, comprising:

an image selector for selecting an image in a best state from a plurality of iris images inputted from an image input device;

an internal edge extractor for extracting an internal edge of the selected iris image;

an external edge extractor for extracting an external edge of the selected iris image;

an iris area normalizer for normalizing an iris area;

a security level applier for securing the iris which increases according to age;

a rotated image processor for processing an inclined iris image;

a feature value extractor for extracting a feature value from the iris image; and

a consistency decider for comparing the extracted feature value with a stored iris code to decide whether the compared feature value and iris code are consistent,

wherein the image selector is operable to scan an inputted iris image in pixel unit by using a mask defined into a second area which is in square shape and a first area configured as the peripheral girth of the second area, calculate a first number of pixels located in the first area for which a luminance value is smaller than a first threshold value and a second number of pixels located in the second area for which a luminance value is greater than a second threshold value, and select an image for which the second number of pixels is minimized.

2. The iris recognition system of claim 1, wherein a central point of the second area of the selected image is designated as an illumination point, and

wherein the internal edge extractor is operable to form one or more circles by grouping calculated plural edge lines as progressing Canny edge detection, search centers of each circle by using a bisection technique for each circle, designate a center nearest to the illumination point as a final pupil central point among centers which are adjacently gathered together, and store an edge constituted by a scale space circle from the final pupil central point as an internal edge, by using a radius of a circle selected as the final pupil central point.

3. The iris recognition system of claim 2, wherein the external edge extractor stores an edge line that constitutes a scale space circle at the final pupil central point as an external edge.

4. The iris recognition system of claim 3, wherein the iris area normalizer is characterized by using iris information existing within a certain radius at the final pupil central point as a feature and by applying an elastic body model so that the elastic body model can prevent transformation caused by movement of the iris, thereby normalizing the iris area by using some areas only of doughnut shapes between the internal edge and the external edge.

5. The iris recognition system of claim 4, wherein the iris area normalizer is configured to divide some areas of the doughnut shapes into at least 3 doughnut shapes in 1, 2, and 3 track sequence which is the nearest doughnut shape sequence to the internal edge at the final pupil central point, and normalize the divided areas.

6. The iris recognition system of claim 5, wherein the security level applier is configured to allocate high weight to the track 1.

7. The iris recognition system of claim 5, wherein the rotated image processor is configured to form the stored iris code into a plurality of formed iris codes with values shifted right and left for an angle inclined right and left by using the stored iris code normalized by the iris area normalizes, store the formed iris codes, 1-dimensionally store the formed iris codes in 1, 2, and 3 track sequence, and compare only the formed iris code stored in track 1.

8. The iris recognition system of claim 1, wherein the feature value extractor is configured to use an LL3 part having the biggest energy value among Daubechies wavelet packets.

9. The iris recognition system of claim 1, further comprising:

a fake iris decider for determining whether an iris is fake by comparing a difference between feature values of irises of two images taken by differentiating the intensity of illumination.

10. An iris recognition method for selecting an iris image by scanning a plurality of iris images inputted from an image input device, in pixel unit by using a mask defined into a second area which is in square shape in the middle part and a first area configured as the peripheral girth of the second area, the method comprising:

determining whether pixels that constitute inputted images exist for which a luminance value of at least one pixel is smaller than a first threshold value;

if so;

scanning luminance values of pixels corresponding to the first area by using the mask, and

determining a first number of pixels for which the corresponding luminance has smaller values than a second threshold value;

determining whether the first number of pixels is greater than 75% of a total number of pixels in the first area of the mask;

if the first number of pixels is greater than 80% of a total number of pixels in the first area of the mask;

scanning luminance values of pixels corresponding to the second area by using the mask, and

counting a second number of pixels for which the corresponding luminance value is greater than a third threshold value;

determining whether the second number of pixels have luminance values within a predetermined range; and

scanning the corresponding image if the second number of pixels have values within the predetermined range.

11. The iris recognition method of claim 10, further comprising:

selecting an image having the smallest pixel in the second number of pixels as an optimum image among various images.

12. An encryption system using an iris code for hashing an iris feature value by using a predetermined pattern, comprising:

an image selector for selecting an optimum image from a plurality of iris images inputted from an image input device;

an internal edge extractor for extracting an internal edge of the selected iris image;

an external edge extractor for extracting an external edge of the selected iris image;

an iris area normalizer for normalizing an iris area;

a security level applier for securing an iris which increases according to age;

a rotated image processor for processing a rotated iris image; and

a feature value extractor for extracting a feature value from the iris image,

wherein the iris area normalizer is configured to divide some areas of doughnut shapes into at least 3 doughnut shapes in 1, 2, and 3 track sequence which is the nearest doughnut shape sequence to the internal edge at the final pupil central point, and normalize the divided areas; and

at least one of the tracks 1, 2, and 3 is used as the predetermined pattern.

13. The encryption system of claim 12, wherein the security level applier is configured to generate a public key and a predetermined pattern (private key) to which random features of the iris itself are applied, by using the 3 tracks.

14. The encryption system of claim 12, wherein the feature value extractor is configured to generate a hash value by selecting an LL3 part of the iris image.

15. The encryption system of claim 12, wherein the feature value extractor is configured to generate an image filter necessary for hash value generation by selecting the LL3 part of the iris image extracted from a spatial domain, a frequency domain, or a scale domain instead of the spatial and frequency domains.

16. The encryption system of claim 12, further comprising:

a fake iris decider for determining whether an iris is fake, wherein the fake iris decider performs Error Correction Code (ECC) encryption by performing an exclusive OR operation for an iris code and an intrinsic iris code hash value to prevent forgery between a stored iris code and a compared iris code.