Iris recognition system using quality metrics

• US 8,280,119 B2
• Filed: 12/05/2008
• Issued: 10/02/2012
• Est. Priority Date: 12/05/2008
• Status: Active Grant

• Current Assignee: Gentex Corporation
• Original Assignee: Honeywell International Inc.
• Inventors: Hamza, Rida
• Primary Examiner(s): LE, BRIAN Q
• Application Number:
  US12/329,346
• Publication Number:
  US 20100142765A1
• Time in Patent Office: 1,397 Days
• Field of Search: 382/103, 382/115, 382/117, 382/118, 382/154, 356/4.03, 356/4.04, 356/4.05, 356/4.06, 356/4.07, 356/18, 356/492, 356/493, 356/496, 356/498
• US Class Current: 382/117
• CPC Class Codes: G06V 10/993 Evaluation of the quality of the acquired pattern
• CPC Class Codes: G06V 40/18 Eye characteristics, e.g. of the iris

Abstract

A system for iris recognition using a set of quality metrics, which may include eye image validation, blur assessment, offset, gazing, obscuration, visibility, and the like. These metrics may be established as quantitative measures which can automatically assess the quality of eye images before they are processed for recognition purposes. Quadrant iris analysis, histograms, map processing enhancements, and multi-band analysis may be used in aiding in the iris recognition approach.

Claims

1. A method for iris recognition comprising:

locating an eye with a camera;

obtaining an image of the eye with the camera;

assessing the image of the eye with a set of image quality metrics with a processor; and

segmenting the iris in the image of the eye with the processor;

wherein the set of image quality metrics comprises;

an offset measurement of the eye in the image of the eye; and

a gaze measurement of the eye in the image of the eye;

wherein a calibration of the segmenting of the iris is determined by the offset and gaze measurements; and

wherein;

if the offset and gaze measurements indicate offset or gaze of the eye in the image of the eye, then the segmenting of the iris is based on no circular calibration; and

if the offset and gaze measurements. indicate no offset or gaze of the eye, then the segmenting of the iris is based on circular calibration.

2. The method of claim 1, further comprising determining a quality score of the image of the eye with the set of image quality metrics with the processor.

3. The method of claim 2, further comprising:

determining a matching score of the image of the eye based on matches of the image of the eye with other images of the eye with the processor; and

establishing an evaluation score based on the quality score and the matching score with the processor.

4. The method of claim 1, wherein the set of image quality metrics further comprises:

a validity measurement of the image of the eye; and

a blur measurement of the image of the eye.

5. A method for iris recognition comprising:

locating an eye with a camera;

obtaining an image of the eye with the camera; and

assessing the image of the eye with a set of image quality metrics with a processor; and

segmenting the iris in the image of the eye with the processor;

wherein the set of image quality metrics comprises;

an offset measurement of the eye in the image of the eye; and

a gaze measurement of the eye in the image of the eye;

wherein a calibration of the segmenting of the iris is determined by the offset and gaze measurements;

wherein the set of image quality metrics further comprises;

an obscuration measurement of the eye in the image of the eye; and

a visibility measurement of the iris of the eye; and

wherein;

if the obscuration measurement of the eye reveals the eye not to be obscured, then a map of an iris of the eye is extracted with the processor; and

a visibility measurement of the iris in the map of the iris is made with the processor.

6. The method of claim 5, further comprising:

coding the iris in a single band code and/or multi-band code with associated weights based on the image quality metrics with the processor; and

matching the single band code and/or multi-band code with other codes of the iris with the processor.

7. The method of claim 6, wherein the set of image quality metrics further comprises:

a validity measurement of the image of the eye; and

a blur measurement of the image of the eye; and

wherein the associated weights based on the image quality metrics used in coding the iris are based at least in part on the blur measurement.

8. A method for iris recognition comprising:

locating an eye with a camera;

obtaining an image of the eye with the camera; and

assessing the image of the eye with a set of image quality metrics with a processor; and

segmenting the iris in the image of the eye with the processor;

wherein the set of image quality metrics comprises;

an offset measurement of the eye in the image of the eye; and

a gaze measurement of the eye in the image of the eye;

wherein a calibration of the segmenting of the iris is determined by the offset and gaze measurements; and

wherein segmenting the iris includes a quadrant iris analysis performed with the processor, the quadrant iris analysis comprising;

finding a first limit in a first quadrant toward a second quadrant, starting from a sclera in a first direction;

finding a second limit in newly defined second quadrant based on the first limit, starting from the sclera at the horizontal axis in a second direction opposite the first direction;

applying an interpolation/mask test to the first and second quadrants;

finding limits in a third quadrant in a direction toward a fourth quadrant;

finding limits in a newly defined fourth quadrant in an opposite direction; and

applying an interpolation/mask test to the third and fourth quadrants.

9. A system for iris recognition using quality metrics, comprising:

an eye image source;

a quality metric mechanism connected to the eye image source;

an iris map extractor connected to the quality metric mechanism and configured to extract a map of an iris; and

a processor configured to eliminate intensity outliers in the map;

wherein clustering pixel intensities of the map is based on their distributed histogram by selectively choosing limits of the iris as represented in the map using associated valleys on the left and right of an iris pixel cluster.

10. A system for iris recognition using quality metrics, comprising:

an eye image source;

a quality metric mechanism connected to the eye image source; and

an iris map extractor connected to the quality metric mechanism and configured to extract a map of an iris;

wherein;

a limited range is imposed to preserve a certain percentage of pixel elements of a prominent pixel cluster which likely represents the iris in the map; and

an element beyond the limited range or the estimated thresholds of the pixel cluster represents noise subject to removal.