Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

On the Feasibility of Creating Iris Periocular Morphed Images

Juan E. Tapia, Sebastian Gonzalez, Daniel Benalcazar, Christoph Busch

TL;DR

The paper addresses the security risks of image-level iris morphing by developing an end-to-end framework that creates periocular iris morphs and evaluates them with a novel SiamIris recognition system. It introduces landmark-based iris morphing, two subject-pair selection strategies, and comprehensive vulnerability analyses using MAP and RMMR, along with a Random Forest/MAD approach. Key findings show that iris-periocular morphs can considerably degrade recognition performance, with radius-based morphs presenting stronger attack potential and posing detection challenges, while the SiamIris system with periocular input yields strong separability. The work highlights practical implications for biometric security, suggesting avenues for robust morph-detection and cross-dataset benchmarking, and sets the stage for comparing with commercial iris systems.

Abstract

In the last few years, face morphing has been shown to be a complex challenge for Face Recognition Systems (FRS). Thus, the evaluation of other biometric modalities such as fingerprint, iris, and others must be explored and evaluated to enhance biometric systems. This work proposes an end-to-end framework to produce iris morphs at the image level, creating morphs from Periocular iris images. This framework considers different stages such as pair subject selection, segmentation, morph creation, and a new iris recognition system. In order to create realistic morphed images, two approaches for subject selection are explored: random selection and similar radius size selection. A vulnerability analysis and a Single Morphing Attack Detection algorithm were also explored. The results show that this approach obtained very realistic images that can confuse conventional iris recognition systems.

On the Feasibility of Creating Iris Periocular Morphed Images

TL;DR

The paper addresses the security risks of image-level iris morphing by developing an end-to-end framework that creates periocular iris morphs and evaluates them with a novel SiamIris recognition system. It introduces landmark-based iris morphing, two subject-pair selection strategies, and comprehensive vulnerability analyses using MAP and RMMR, along with a Random Forest/MAD approach. Key findings show that iris-periocular morphs can considerably degrade recognition performance, with radius-based morphs presenting stronger attack potential and posing detection challenges, while the SiamIris system with periocular input yields strong separability. The work highlights practical implications for biometric security, suggesting avenues for robust morph-detection and cross-dataset benchmarking, and sets the stage for comparing with commercial iris systems.

Abstract

In the last few years, face morphing has been shown to be a complex challenge for Face Recognition Systems (FRS). Thus, the evaluation of other biometric modalities such as fingerprint, iris, and others must be explored and evaluated to enhance biometric systems. This work proposes an end-to-end framework to produce iris morphs at the image level, creating morphs from Periocular iris images. This framework considers different stages such as pair subject selection, segmentation, morph creation, and a new iris recognition system. In order to create realistic morphed images, two approaches for subject selection are explored: random selection and similar radius size selection. A vulnerability analysis and a Single Morphing Attack Detection algorithm were also explored. The results show that this approach obtained very realistic images that can confuse conventional iris recognition systems.
Paper Structure (21 sections, 4 equations, 9 figures, 8 tables)

This paper contains 21 sections, 4 equations, 9 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Iris Morphing
  4. Segmentation
  5. Iris Recognition
  6. Metrics
  7. d' Value
  8. Morphing Attack Classification Error Rate
  9. Morph Vulnerability Analysis
  10. Dataset
  11. Methods
  12. Iris Recognition
  13. Segmentation
  14. Morph
  15. Pairs selection
  16. ...and 6 more sections

Figures (9)

  • Figure 1: Top: Example image from CASIA-IrisV4-Lamp. Bottom: UND databases. The grey border is an artefact produced by the capture device.
  • Figure 2: SiamIris Architecture. This approach predicts whether two irises come from the same or different subjects. This iris recognition system can be processed as periocular images or rubber-sheet images. For example a) uses periocular iris images as inputs, whereas in b) rubber-sheet images are used as inputs.
  • Figure 3: Image pre-processing steps. a) Input image. b) Iris mask predicted by DenseNet10. c) Iris and pupil localization using the mixed algorithm. d) Isolated iris. e) Rubber-sheet model. Blue lines represent the pupil-iris boundary, while red lines represent the iris-sclera boundary. The green arrows in d) and e) represent the direction in which the iris was unwrapped.
  • Figure 4: Iris morphing image generation process.
  • Figure 5: Example of periocular morph images. Top: UND bona fide periocular images. Middle: Morph images using random selection. Bottom: Morph images using the radii pupil size method.
  • ...and 4 more figures