Deep Learning for Iris Recognition A Survey

2025-04-26 0 0 1.79MB 35 页 10玖币

侵权投诉

Deep Learning for Iris Recognition: A Survey

KIEN NGUYEN, Queensland University of Technology, Australia

HUGO PROENÇA, University of Beira Interior, IT: Instituto de Telecomunicações, Portugal

FERNANDO ALONSO-FERNANDEZ, Halmstad University, Sweden

ABSTRACT

In this survey, we provide a comprehensive review of more than 200 papers, technical reports, and GitHub

repositories published over the last 10 years on the recent developments of deep learning techniques for iris

recognition, covering broad topics on algorithm designs, open-source tools, open challenges, and emerging

research. First, we conduct a comprehensive analysis of deep learning techniques developed for two main

sub-tasks in iris biometrics: segmentation and recognition. Second, we focus on deep learning techniques for

the robustness of iris recognition systems against presentation attacks and via human-machine pairing. Third,

we delve deep into deep learning techniques for forensic application, especially in post-mortem iris recognition.

Fourth, we review open-source resources and tools in deep learning techniques for iris recognition. Finally,

we highlight the technical challenges, emerging research trends, and outlook for the future of deep learning

in iris recognition.

CCS Concepts: •Security and privacy →Biometrics.

Additional Key Words and Phrases: Iris Recognition, Deep Learning, Neural Networks

ACM Reference Format:

Kien Nguyen, Hugo Proença, and Fernando Alonso-Fernandez. 2022. Deep Learning for Iris Recognition: A

Survey. 1, 1 (October 2022), 35 pages. https://doi.org/10.1145/nnnnnnn.nnnnnnn

1 INTRODUCTION

The human iris is a sight organ that controls the amount of light reaching the retina, by changing

the size of the pupil. The texture of the iris is fully developed before birth, its minutiae do not depend

on genotype, it stays relatively stable across lifetime (except for disease- and normal aging-related

biological changes), and it may even be used for forensic identication shortly after subject’s

death [36, 110, 170].

In terms of its information theory-related properties, the iris texture has an extremely high

randotypic randomness, and is stable (permanent) over time, providing an exceptionally high

entropy per mm.

that justies its higher discriminating power, when compared to other biometric

modalities (e.g., face or ngerprint). The iris’ collectability is another feature of interest and has been

the subject of discussion over the last years: while it can be acquired using commercial o-the-shelf

(COTS) hardware, either handheld or stationary, data can be even collected from at-a-distance, up to

tens of meters away from the subjects [

111

]. Even though commercial visible-light (RGB) cameras

are able to image the iris, the near infrared-based (NIR) sensing dominates in most applications, due

Authors’ addresses: Kien Nguyen, Queensland University of Technology, Australia, nguyentk@qut.edu.au; Hugo Proença,

University of Beira Interior, IT: Instituto de Telecomunicações, Portugal, hugomcp@di.ubi.pt; Fernando Alonso-Fernandez,

Halmstad University, Sweden, feralo@hh.se.

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee

provided that copies are not made or distributed for prot or commercial advantage and that copies bear this notice and

the full citation on the rst page. Copyrights for components of this work owned by others than ACM must be honored.

Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires

prior specic permission and/or a fee. Request permissions from permissions@acm.org.

XXXX-XXXX/2022/10-ART $15.00

https://doi.org/10.1145/nnnnnnn.nnnnnnn

, Vol. 1, No. 1, Article . Publication date: October 2022.

arXiv:2210.05866v1 [cs.CV] 12 Oct 2022

2 K. Nguyen, H. Proença, F. Alonso-Fernandez

to a better visibility of iris texture for darker eyes, rich in melanin pigment, which is characterized

by lower light absorption in NIR spectrum compared to shorter wavelengths. In addition, NIR

wavelengths are barely perceivable by the human eye, which augment users’ comfort, and avoids

pupil contraction/dilation that would appear under visible light.

A seminal work by John Daugman brought to the community the Gabor ltering-based approach

that became the dominant approach for iris recognition [

]. Even though subsequent

solutions to iris image encoding and matching appeared, the IrisCodes approach is still dominant due

to its ability to eectively search in massive databases with a minimal probability of false matches,

at extreme time performance. By considering binary words, pairs of signatures are matched using

XOR parallel-bit logic at lightening speed, enabling millions of comparisons/second per processing

core. Also, most of the methods that outperformed the original techniques in terms of eectiveness

do not work under the one-shot learning paradigm, assume multiple observations of each class to

obtain appropriate decision boundaries, and - most importantly - have encoding/matching steps

with time complexity that forbid their use in large environments (in particular, for all-against-all

settings).

In short, Daugman’s algorithm encodes the iris image into a binary sequence of 2,048 bits by

ltering the iris image with a family of Gabor kernels. The varying pupil size is rectied by the

Cartesian-to-polar coordinate system transformation, to end up with an image representation of

canonical size, guarantying identical structure of the iris code independently of the iris and pupil

size. This makes possible to use the Hamming Distance (HD) to measure the similarity between two

iris codes [

]. Its low false match rate at acceptable false non-match rates is the key factor behind

the success of global-scale iris recognition installments, such as the national person identication

and border security program Aadhaar program in India (with over 1.2 billion pairs of irises enrolled)

[

174

], the Homeland Advanced Recognition Technology (HART) in the US (up to 500 million

identities) [

128

], or the NEXUS system, designed to speed up border crossings for low-risk and

pre-approved travelers moving between Canada and the US.

Deep learning-based methods, in particular using various Convolutional Neural Network archi-

tectures, have been driving remarkable improvements in many computer vision applications over

the last decade. In terms of biometrics technologies, it’s not surprising that iris recognition has

also seen an increasing adoption of purely data-driven approaches at all stages of the recognition

pipeline: from preprocessing (such as o-axis gaze correction), segmentation, encoding to matching.

Interestingly, however, the impact of deep learning on the various stages of iris recognition pipeline

is uneven. One of the primary goals of this survey paper is to assess where deep learning helped

in achieving highly performance and more secure systems, and which procedures did not benet

from more complex modeling.

The remainder of the paper is structured as follows. Section 2 and 3 review the application of deep

learning in two main stages of the recognition pipeline: segmentation and recognition (encoding

and comparison). Section 4 and 5 analyze the state of the art of deep learning-based approaches in

two applications: Presentation Attack Detection (PAD) and Forensic. Section 6 investigates how

human and machine can pair to improve deep learning based iris recognition. Section 7 focuses on

approaches in less controlled environments of iris and periocular analysis. Section 8 reviews public

resources and tools available in the deep learning based iris recognition domain. Section 9 focuses

on the future of deep learning for iris recognition with discussion on emerging research directions

in dierent aspects of iris analysis. The paper in concluded in Section 10.

2 DEEP LEARNING-BASED IRIS SEGMENTATION

The segmentation of the iris is seen as an extremely challenging problem. As illustrated in Fig. 1,

segmenting the iris involves essentially three tasks: detect and parameterize the inner (pupillary)

, Vol. 1, No. 1, Article . Publication date: October 2022.

Deep Learning for Iris Recognition: A Survey 3

and outer (scleric) biological boundaries of the iris and also to locally discriminate between the

noise-free/noisy regions inside the iris ring, which should be subsequently used in the feature

encoding and matching processes.

This problem has motivated numerous research works for decades. From the pioneering integro-

dierential operator [

] up to subsequent handcrafted techniques based in active contours and

neural networks (e.g., [

], [

133

], [

147

] and [

175

]) a long road has been traveled in this problem.

Regardless an obvious evolution in the eectiveness of such techniques, they all face particular

diculties in case of heavily degraded data. Images are frequently motion-blurred, poor focused,

partially occluded and o-angle. Additionally, in case of visible light data, severe reections from

the environments surrounding the subjects are visible, and even augment the diculties of the

segmentation task.

Recently, as in many other computer vision tasks, DL-based frameworks have been advocated as

providing consistent advances over the state-of-the-art for the iris segmentation problem, with

numerous models being proposed. A cohesive perspective of the most relevant recent DL-based

methods is given in Table 1, with the techniques appearing in chronographic (and then alphabetical)

order. The type of data each model aims to handle is given in the ”Data” column, along with

the datasets where the corresponding experiments were carried out and a summary of the main

characteristics of each proposal (”Features” column). Here, considering that models were empirically

validated in completely heterogeneous ways and using very dierent metrics, we decided not to

include the summary performance of each model/solution.

Scleric Boundary Parameterization 2

Noise-free Texture Detection 3

Pupillary Boundary Parameterization 1

Iris Segmentation Main Tasks

Dimensionless Noise-Free

Representation

1+ + 3

Fig. 1. Three main tasks typically associated to iris segmentation: 1) parameterization of the pupillary (inner)

boundary; 2) parameterization of the scleric (outer) boundary; and 3) discrimination between the unoccluded

(noise-free) and occluded (noisy) regions inside the iris ring. Such pieces of information are further used to

obtain dimensionless polar representations of the iris texture, where feature extraction methods typically

operate.

Schlett et al. [

144

] provided a multi-spectral analysis to improve iris segmentation accuracy

in visible wavelengths by preprocessing data before the actual segmentation phase, extracting

multiple spectral components in form of RGB color channels. Even though this approach does

propose a DL-based framework, the dierent versions of the input could be easily used to feed

DL-based models, and augment the robustness to non-ideal data. Chen et al. [

] used CNNs

that include dense blocks, referred to as a dense-fully convolutional network (DFCN), where the

encoder part consists of dense blocks, and the decoder counterpart obtains the segmentation

masks via transpose convolutions. Hofbauer et al. [

] parameterize the iris boundaries based on

segmentation maps yielding from a CNN, using a a cascaded architecture with four ReneNet

units, each directly connecting to one Residual net. Huynh et al. [

] discriminate between three

, Vol. 1, No. 1, Article . Publication date: October 2022.

4 K. Nguyen, H. Proença, F. Alonso-Fernandez

distinct eye regions with a DL model, and removes incorrect areas with heuristic lters. The

proposed architecture is based on the encoder-decoder model, with depth-wise convolutions used

to reduce the computational cost. Roughly at the same time, Li et al. [

] described the Interleaved

Residual U-Net model for semantic segmentation and iris mask synthesis. In this work, unsupervised

techniques (K-means clustering) were used to create intermediary pictorial representations of the

ocular region, from where saliency points deemed to belong to the iris boundaries were found.

Kerrigan et al. [

] assessed the performance of four dierent convolutional architectures designed

for semantic segmentation. Two of these models were based in dilated convolutions, as proposed

by Yu and Koltun [

188

]. Wu and Zhao [

186

] described the Dense U-Net model, that combines dense

layers to the U-Net network. The idea is to take advantage of the reduced set of parameters of

the dense U-Net, while keeping the semantic segmentation capabilities of U-Net. The proposed

model integrates dense connectivity into U-Net contraction and expansion paths. Compared with

traditional CNNs, this model is claimed to reduce learning redundancy and enhance information

ow, while keeping controlled the number of parameters of the model. Wei et al. [

205

] suggested

to perform dilated convolutions, which is claimed to obtain more consistent global features. In

this setting, convolutional kernels are not continuous, with zero-values being articially inserted

between each non-zero position, increasing the receptive eld without augmenting the number of

parameters of the model.

More recently, Ganeva and Myasnikov [

] compared the eectiveness of three convolutional

neural network architectures (U-Net, LinkNet, and FC- DenseNet), determining the optimal pa-

rameterization for each one. Jalilian et al. [

] introduced a scheme to compensate for texture

deformations caused by the o-angle distortions, re-projecting the o-angle images back to frontal

view. The used architecture is a variant of ReneNet [

], which provides high-resolution prediction,

while preserving the boundary information (required for parameterization purposes).

The idea of interactive learning for iris segmentation was suggested by Sardar et al. [

142

describing an interactive variant of U-Net that includes Squeeze Expand modules. Trokielewicz

et al. [

172

] used DL-based iris segmentation models to extract highly irregular iris texture areas

in post-mortem iris images. They used a pre-trained SegNet model, ne-tuned with a database

of cadaver iris images. Wang et al. [

178

] (further extended in [

179

]) described a lightweight deep

convolutional neural network specically designed for iris segmentation of degraded images

acquired by handheld devices. The proposed approach jointly obtains the segmentation mask and

parameterized pupillary/limbic boundaries of the iris.

Observing that edge-based information is extremely sensitive to be obtained in degraded data,

Li et al. [

] presented an hybrid method that combines edge-based information to deep learning

frameworks. A compacted Faster R-CNN-like architecture was used to roughly detect the eye and

dene the initial region of interest, from where the pupil is further located using a Gaussian mixture

model. Wang et al. [

184

] trained a deep convolutional neural network(DCNN) that automatically

extracts the iris and pupil pixels of each eye from input images. This work combines the power of

U-Net and SqueezeNet to obtain a compact CNN suitable for real time mobile applications. Finally,

Wang et al. [

176

] parameterize both the iris mask and the inner/outer iris boundaries jointly, by

actively modeling such information into a unied multi-task network.

A nal word is given to segmentation-less techniques. Assuming that the accurate segmentation of

the iris boundaries is one of the hardest phases of the whole recognition chain and the main source

for recognition errors, some recent works have been proposing to perform biometrics recognition

in non-segmented or roughly segmented data [

132

][

135

]. Here, the idea is to use the remarkable

discriminating power of DL-frameworks to perceive the agreeing patterns between pairs of images,

even on such segmentation-less representations.

, Vol. 1, No. 1, Article . Publication date: October 2022.

Deep Learning for Iris Recognition: A Survey 5

Table 1. Cohesive comparison of the most relevant DL-based iris segmentation methods (NIR: near-infrared;

VW: visible wavelength). Methods are listed in chronological (and then alphabetical) order.

Method Year Data Datasets Features

NIR VW

Schlett et al. [144] 2018 ✗ ✓ MobBIO Preprocessing (combines dierent possibili-

ties of the input RGB channels)

Trokielewicz and

Czajka [166]

2018 ✓ ✓ Warsaw-Post-Mortem v1.0 Fine-tuned CNN (SegNet)

Chen et al. [22] 2019 ✓ ✓ CASIA-Irisv4-Interval, IITD, UBIRIS.v2 Dense CNN

Hofbauer et

al. [72]

2019 ✓ ✗ IITD, CASIA-Irisv4-Interval, ND-Iris-0405 Cascaded architecture of four ReneNet,

each connecting to one Residual net

Huynh et al. [76] 2019 ✓ ✗ OpenEDS MobileNetV2 + heuristic ltering postproc.

Li et al. [7] 2019 ✓ ✗ CASIA-Iris-Thousand Faster-R-CNN (ROI detection)

Kerrigan et al. [85] 2019 ✓ ✓ CASIA-Irisv4-Interval, BioSec, ND-Iris-

0405, UBIRIS.v2, Warsaw-Post-Mortem

v2.0, ND-TWINS-2009-2010

Resent + Segnet (with dilated convolutions)

Wu and Zhao [186] 2019 ✓ ✓ CASIA-Irisv4-Interval, UBIRIS.v2 Dense-U-Net (dense layers + U-Net)

Wei et al. [205] 2019 ✓ ✓ CASIA-Iris4-Interval, ND-IRIS-0405,

UBIRIS.v2

U-Net with dilated convolutions

Fang and

Czajka [50]

2020 ✓ ✓ ND-Iris-0405, CASIA, BATH, BioSec,

UBIRIS, Warsaw-Post-Mortem v1.0 & v2.0

Fine-tuned CC-Net [106]

Ganeva and Myas-

nikov [55]

2020 ✓ ✗ MMU U-Net, LinkNet, and FC-DenseNet (perfor-

mance comparison)

Jalilian et al. [79] 2020 ✓ ✗ ReneNet + morphological postprocessing

Sardar et al. [142] 2020 ✓ ✓ CASIA-Irisv4-Interval, IITD, UBIRIS.v2 Squeeze-Expand module + active learning

(interactive segmentation)

Trokielewicz et

al. [172]

2020 ✓ ✓ ND-Iris-0405, CASIA, BATH, BioSec,

UBIRIS, Warsaw-Post-Mortem v1.0 & v2.0

Fined-tuned SegNet [9]

Wang et al [178] 2020 ✓ ✓ CASIA-Iris-M1-S1/S2/S3, MICHE-I Hourglass network

Wang et al. [176] 2020 ✓ ✓ CASIA-v4-Distance, UBIRIS.v2, MICHE-I U-Net + multi-task attention net + postproc.

(probabilistic masks priors & thresholding)

Li et al. [94] 2021 ✓ ✗ CASIA-Iris-Thousand IRU-Net network

Wang et al. [184] 2021 ✗ ✓ Online Video Streams and Internet Videos U-Net and Squeezenet to iris segmentation

and detect eye closure

Kuehlkamp

et al. [91]

2022 ✓ ✓ ND-Iris-0405, CASIA, BATH, BioSec,

UBIRIS, Warsaw-Post-Mortem v2.0

Fined-tuning of Mask-RCNN architecture

3 DEEP LEARNING-BASED IRIS RECOGNITION

3.1 Deep Learning Models as a Feature Extractor

As illustrated in Fig. 2, the idea here is to analyze a dimensionless representation of the iris data and

produce a feature vector that lies in a hyperspace (embedding) where recognition is carried out.

In this context, Boyd et el. [15] explored ve dierent sets of weights for the popular ResNet50

architecture to test if iris-specic feature extractors perform better than models trained for general

tasks. Minaee et al. [

105

] studied the application of deep features extracted from VGG-Net for

iris recognition, having authors observed that the resulting features can be well transferred to

biometric recognition. Luo et al. [

102

] described a DL model with spatial attention and channel

attention mechanisms, that are directly inserted into the feature extraction module. Also, a co-

attention mechanism adaptively fuses features to obtain representative iris-periocular features.

Hafner et al. [

] adapted the classical Daugman’s pipeline, using convolutional neural networks

, Vol. 1, No. 1, Article . Publication date: October 2022.

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

DeepLearningforIrisRecognition:ASurveyKIENNGUYEN,QueenslandUniversityofTechnology,AustraliaHUGOPROENÇA,UniversityofBeiraInterior,IT:InstitutodeTelecomunicações,PortugalFERNANDOALONSO-FERNANDEZ,HalmstadUniversity,SwedenABSTRACTInthissurvey,weprovideacomprehensivereviewofmorethan200papers,technicalrep...

展开>> 收起<<

Deep Learning for Iris Recognition A Survey.pdf

共35页,预览5页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Deep Learning for Iris Recognition A Survey

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: