Using Non-Orthogonal Iris Images for Iris Recognition

Abstract

The iris is the colored portion of the eye that surrounds the pupil and controls the amount of light that can enter the eye. The variations within the patterns of the iris are unique between eyes, which allows for accurate identification of an individual. Current commercial iris recognition algorithms require an orthogonal image of the eye (subject is looking directly into a camera) to find circular inner (pupillary) and outer (limbic) boundaries of the iris. If the subject is looking away from the camera (non-orthogonal), the pupillary and limbic boundaries appear elliptical, which a commercial system may be unable to process. This elliptical appearance also reduces the amount of information that is available in the image used for recognition. These are major challenges in non-orthogonal iris recognition. This research addressed these issues and provided a means to perform non-orthogonal iris recognition. All objectives set forth at the start of this project were accomplished. The first major objective of this project was to construct a database of non-orthogonal iris images for algorithm development and testing. A collection station was built that allows for the capture of iris images at 0 degrees (orthogonal), 15 degrees, 30 degrees, and 45 degrees. During a single collection on an individual, nine images were collected at each angle for each eye. Images of approximately 90 irises were taken, with 36 images collected per eye. Sixty irises were evaluated twice, resulting in a total of almost 7100 images in the database. The second major objective involved modifying the Naval Academy's one-dimensional iris recognition algorithm so it could process non-orthogonal iris images. An elliptical-to-circular (affine) transformation was applied to the nonorthogonal images to create circular boundaries. This permitted the algorithm to be run as designed, with this modified algorithm used in the recognition testing phase of the project.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 05, 2006
Accession Number
ADA453040

Entities

People

  • Ruth M. Gaunt

Organizations

  • United States Naval Academy

Tags

Communities of Interest

  • Air Platforms
  • Sensors

DTIC Thesaurus Topics

  • Accuracy
  • Algorithms
  • Biometric Security
  • Biometrics
  • Boundaries
  • Databases
  • Detection
  • Electrical Engineering
  • Engineering
  • Graphical User Interface
  • Identification
  • Image Processing
  • Recognition
  • Shape
  • Signal Processing
  • United States Naval Academy
  • User Interface

Readers

  • Computer Vision.
  • Vision Science/Vision Psychology/Cognitive Neuroscience.