The overlapping effect and fusion protocols of data augmentation techniques in iris PAD View Full Text


Ontology type: schema:ScholarlyArticle      Open Access: True


Article Info

DATE

2021-11-26

AUTHORS

Meiling Fang, Naser Damer, Fadi Boutros, Florian Kirchbuchner, Arjan Kuijper

ABSTRACT

Iris Presentation Attack Detection (PAD) algorithms address the vulnerability of iris recognition systems to presentation attacks. With the great success of deep learning methods in various computer vision fields, neural network-based iris PAD algorithms emerged. However, most PAD networks suffer from overfitting due to insufficient iris data variability. Therefore, we explore the impact of various data augmentation techniques on performance and the generalizability of iris PAD. We apply several data augmentation methods to generate variability, such as shift, rotation, and brightness. We provide in-depth analyses of the overlapping effect of these methods on performance. In addition to these widely used augmentation techniques, we also propose an augmentation selection protocol based on the assumption that various augmentation techniques contribute differently to the PAD performance. Moreover, two fusion methods are performed for more comparisons: the strategy-level and the score-level combination. We demonstrate experiments on two fine-tuned models and one trained from the scratch network and perform on the datasets in the Iris-LivDet-2017 competition designed for generalizability evaluation. Our experimental results show that augmentation methods improve iris PAD performance in many cases. Our least overlap-based augmentation selection protocol achieves the lower error rates for two networks. Besides, the shift augmentation strategy also exceeds state-of-the-art (SoTA) algorithms on the Clarkson and IIITD-WVU datasets. More... »

PAGES

8

References to SciGraph publications

  • 2017-04-21. Iris recognition with tunable filter bank based feature in MULTIMEDIA TOOLS AND APPLICATIONS
  • 2019-07-06. A survey on Image Data Augmentation for Deep Learning in JOURNAL OF BIG DATA
  • 2020-09-22. Iris presentation attack detection based on best-k feature selection from YOLO inspired RoI in NEURAL COMPUTING AND APPLICATIONS
  • Identifiers

    URI

    http://scigraph.springernature.com/pub.10.1007/s00138-021-01256-9

    DOI

    http://dx.doi.org/10.1007/s00138-021-01256-9

    DIMENSIONS

    https://app.dimensions.ai/details/publication/pub.1143312965


    Indexing Status Check whether this publication has been indexed by Scopus and Web Of Science using the SN Indexing Status Tool
    Incoming Citations Browse incoming citations for this publication using opencitations.net

    JSON-LD is the canonical representation for SciGraph data.

    TIP: You can open this SciGraph record using an external JSON-LD service: JSON-LD Playground Google SDTT

    [
      {
        "@context": "https://springernature.github.io/scigraph/jsonld/sgcontext.json", 
        "about": [
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/08", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Information and Computing Sciences", 
            "type": "DefinedTerm"
          }, 
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/0801", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Artificial Intelligence and Image Processing", 
            "type": "DefinedTerm"
          }
        ], 
        "author": [
          {
            "affiliation": {
              "alternateName": "Mathematical and Applied Visual Computing, TU Darmstadt, Darmstadt, Germany", 
              "id": "http://www.grid.ac/institutes/grid.6546.1", 
              "name": [
                "Fraunhofer Institute for Computer Graphics Research IGD, Darmstadt, Germany", 
                "Mathematical and Applied Visual Computing, TU Darmstadt, Darmstadt, Germany"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Fang", 
            "givenName": "Meiling", 
            "id": "sg:person.015017674621.27", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.015017674621.27"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "Mathematical and Applied Visual Computing, TU Darmstadt, Darmstadt, Germany", 
              "id": "http://www.grid.ac/institutes/grid.6546.1", 
              "name": [
                "Fraunhofer Institute for Computer Graphics Research IGD, Darmstadt, Germany", 
                "Mathematical and Applied Visual Computing, TU Darmstadt, Darmstadt, Germany"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Damer", 
            "givenName": "Naser", 
            "id": "sg:person.011045327242.12", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.011045327242.12"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "Mathematical and Applied Visual Computing, TU Darmstadt, Darmstadt, Germany", 
              "id": "http://www.grid.ac/institutes/grid.6546.1", 
              "name": [
                "Fraunhofer Institute for Computer Graphics Research IGD, Darmstadt, Germany", 
                "Mathematical and Applied Visual Computing, TU Darmstadt, Darmstadt, Germany"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Boutros", 
            "givenName": "Fadi", 
            "id": "sg:person.012172532457.60", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.012172532457.60"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "Fraunhofer Institute for Computer Graphics Research IGD, Darmstadt, Germany", 
              "id": "http://www.grid.ac/institutes/grid.461618.c", 
              "name": [
                "Fraunhofer Institute for Computer Graphics Research IGD, Darmstadt, Germany"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Kirchbuchner", 
            "givenName": "Florian", 
            "id": "sg:person.07371035657.43", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.07371035657.43"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "Mathematical and Applied Visual Computing, TU Darmstadt, Darmstadt, Germany", 
              "id": "http://www.grid.ac/institutes/grid.6546.1", 
              "name": [
                "Fraunhofer Institute for Computer Graphics Research IGD, Darmstadt, Germany", 
                "Mathematical and Applied Visual Computing, TU Darmstadt, Darmstadt, Germany"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Kuijper", 
            "givenName": "Arjan", 
            "id": "sg:person.015632044271.26", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.015632044271.26"
            ], 
            "type": "Person"
          }
        ], 
        "citation": [
          {
            "id": "sg:pub.10.1186/s40537-019-0197-0", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1117799804", 
              "https://doi.org/10.1186/s40537-019-0197-0"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/s00521-020-05342-3", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1131099745", 
              "https://doi.org/10.1007/s00521-020-05342-3"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/s11042-017-4668-z", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1085027903", 
              "https://doi.org/10.1007/s11042-017-4668-z"
            ], 
            "type": "CreativeWork"
          }
        ], 
        "datePublished": "2021-11-26", 
        "datePublishedReg": "2021-11-26", 
        "description": "Iris Presentation Attack Detection (PAD) algorithms address the vulnerability of iris recognition systems to presentation attacks. With the great success of deep learning methods in various computer vision fields, neural network-based iris PAD algorithms emerged. However, most PAD networks suffer from overfitting due to insufficient iris data variability. Therefore, we explore the impact of various data augmentation techniques on performance and the generalizability of iris PAD. We apply several data augmentation methods to generate variability, such as shift, rotation, and brightness. We provide in-depth analyses of the overlapping effect of these methods on performance. In addition to these widely used augmentation techniques, we also propose an augmentation selection protocol based on the assumption that various augmentation techniques contribute differently to the PAD performance. Moreover, two fusion methods are performed for more comparisons: the strategy-level and the score-level combination. We demonstrate experiments on two fine-tuned models and one trained from the scratch network and perform on the datasets in the Iris-LivDet-2017 competition designed for generalizability evaluation. Our experimental results show that augmentation methods improve iris PAD performance in many cases. Our least overlap-based augmentation selection protocol achieves the lower error rates for two networks. Besides, the shift augmentation strategy also exceeds state-of-the-art (SoTA) algorithms on the Clarkson and IIITD-WVU datasets.", 
        "genre": "article", 
        "id": "sg:pub.10.1007/s00138-021-01256-9", 
        "inLanguage": "en", 
        "isAccessibleForFree": true, 
        "isPartOf": [
          {
            "id": "sg:journal.1045266", 
            "issn": [
              "0932-8092", 
              "1432-1769"
            ], 
            "name": "Machine Vision and Applications", 
            "publisher": "Springer Nature", 
            "type": "Periodical"
          }, 
          {
            "issueNumber": "1", 
            "type": "PublicationIssue"
          }, 
          {
            "type": "PublicationVolume", 
            "volumeNumber": "33"
          }
        ], 
        "keywords": [
          "data augmentation techniques", 
          "augmentation techniques", 
          "augmentation method", 
          "presentation attack detection algorithms", 
          "computer vision field", 
          "attack detection algorithm", 
          "deep learning methods", 
          "score-level combination", 
          "data augmentation method", 
          "iris recognition system", 
          "low error rate", 
          "presentation attacks", 
          "vision field", 
          "recognition system", 
          "art algorithms", 
          "learning method", 
          "PAD algorithms", 
          "detection algorithm", 
          "fusion method", 
          "selection protocol", 
          "algorithm", 
          "network", 
          "error rate", 
          "great success", 
          "augmentation strategies", 
          "experimental results", 
          "dataset", 
          "fusion protocol", 
          "protocol", 
          "performance", 
          "depth analysis", 
          "data variability", 
          "technique", 
          "attacks", 
          "method", 
          "pad performance", 
          "more comparisons", 
          "system", 
          "vulnerability", 
          "model", 
          "experiments", 
          "overlapping effect", 
          "evaluation", 
          "generalizability", 
          "success", 
          "strategies", 
          "field", 
          "assumption", 
          "brightness", 
          "results", 
          "state", 
          "Clarkson", 
          "combination", 
          "competition", 
          "analysis", 
          "comparison", 
          "cases", 
          "addition", 
          "impact", 
          "rotation", 
          "variability", 
          "pads", 
          "rate", 
          "shift", 
          "effect", 
          "Iris Presentation Attack Detection (PAD) algorithms", 
          "neural network-based iris PAD algorithms", 
          "network-based iris PAD algorithms", 
          "iris PAD algorithms", 
          "most PAD networks", 
          "PAD networks", 
          "insufficient iris data variability", 
          "iris data variability", 
          "iris PAD", 
          "augmentation selection protocol", 
          "scratch network", 
          "Iris-LivDet", 
          "generalizability evaluation", 
          "iris PAD performance", 
          "least overlap-based augmentation selection protocol", 
          "overlap-based augmentation selection protocol", 
          "shift augmentation strategy", 
          "IIITD-WVU datasets"
        ], 
        "name": "The overlapping effect and fusion protocols of data augmentation techniques in iris PAD", 
        "pagination": "8", 
        "productId": [
          {
            "name": "dimensions_id", 
            "type": "PropertyValue", 
            "value": [
              "pub.1143312965"
            ]
          }, 
          {
            "name": "doi", 
            "type": "PropertyValue", 
            "value": [
              "10.1007/s00138-021-01256-9"
            ]
          }
        ], 
        "sameAs": [
          "https://doi.org/10.1007/s00138-021-01256-9", 
          "https://app.dimensions.ai/details/publication/pub.1143312965"
        ], 
        "sdDataset": "articles", 
        "sdDatePublished": "2022-01-01T19:01", 
        "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
        "sdPublisher": {
          "name": "Springer Nature - SN SciGraph project", 
          "type": "Organization"
        }, 
        "sdSource": "s3://com-springernature-scigraph/baseset/20220101/entities/gbq_results/article/article_895.jsonl", 
        "type": "ScholarlyArticle", 
        "url": "https://doi.org/10.1007/s00138-021-01256-9"
      }
    ]
     

    Download the RDF metadata as:  json-ld nt turtle xml License info

    HOW TO GET THIS DATA PROGRAMMATICALLY:

    JSON-LD is a popular format for linked data which is fully compatible with JSON.

    curl -H 'Accept: application/ld+json' 'https://scigraph.springernature.com/pub.10.1007/s00138-021-01256-9'

    N-Triples is a line-based linked data format ideal for batch operations.

    curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/s00138-021-01256-9'

    Turtle is a human-readable linked data format.

    curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/s00138-021-01256-9'

    RDF/XML is a standard XML format for linked data.

    curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/s00138-021-01256-9'


     

    This table displays all metadata directly associated to this object as RDF triples.

    185 TRIPLES      22 PREDICATES      111 URIs      100 LITERALS      6 BLANK NODES

    Subject Predicate Object
    1 sg:pub.10.1007/s00138-021-01256-9 schema:about anzsrc-for:08
    2 anzsrc-for:0801
    3 schema:author N4320566f995a49e9a9337dd909244d86
    4 schema:citation sg:pub.10.1007/s00521-020-05342-3
    5 sg:pub.10.1007/s11042-017-4668-z
    6 sg:pub.10.1186/s40537-019-0197-0
    7 schema:datePublished 2021-11-26
    8 schema:datePublishedReg 2021-11-26
    9 schema:description Iris Presentation Attack Detection (PAD) algorithms address the vulnerability of iris recognition systems to presentation attacks. With the great success of deep learning methods in various computer vision fields, neural network-based iris PAD algorithms emerged. However, most PAD networks suffer from overfitting due to insufficient iris data variability. Therefore, we explore the impact of various data augmentation techniques on performance and the generalizability of iris PAD. We apply several data augmentation methods to generate variability, such as shift, rotation, and brightness. We provide in-depth analyses of the overlapping effect of these methods on performance. In addition to these widely used augmentation techniques, we also propose an augmentation selection protocol based on the assumption that various augmentation techniques contribute differently to the PAD performance. Moreover, two fusion methods are performed for more comparisons: the strategy-level and the score-level combination. We demonstrate experiments on two fine-tuned models and one trained from the scratch network and perform on the datasets in the Iris-LivDet-2017 competition designed for generalizability evaluation. Our experimental results show that augmentation methods improve iris PAD performance in many cases. Our least overlap-based augmentation selection protocol achieves the lower error rates for two networks. Besides, the shift augmentation strategy also exceeds state-of-the-art (SoTA) algorithms on the Clarkson and IIITD-WVU datasets.
    10 schema:genre article
    11 schema:inLanguage en
    12 schema:isAccessibleForFree true
    13 schema:isPartOf N7385d04271a34d43a6766dbc9e4a6528
    14 Nb74230a12272428a9d5b1a8878bf94f9
    15 sg:journal.1045266
    16 schema:keywords Clarkson
    17 IIITD-WVU datasets
    18 Iris Presentation Attack Detection (PAD) algorithms
    19 Iris-LivDet
    20 PAD algorithms
    21 PAD networks
    22 addition
    23 algorithm
    24 analysis
    25 art algorithms
    26 assumption
    27 attack detection algorithm
    28 attacks
    29 augmentation method
    30 augmentation selection protocol
    31 augmentation strategies
    32 augmentation techniques
    33 brightness
    34 cases
    35 combination
    36 comparison
    37 competition
    38 computer vision field
    39 data augmentation method
    40 data augmentation techniques
    41 data variability
    42 dataset
    43 deep learning methods
    44 depth analysis
    45 detection algorithm
    46 effect
    47 error rate
    48 evaluation
    49 experimental results
    50 experiments
    51 field
    52 fusion method
    53 fusion protocol
    54 generalizability
    55 generalizability evaluation
    56 great success
    57 impact
    58 insufficient iris data variability
    59 iris PAD
    60 iris PAD algorithms
    61 iris PAD performance
    62 iris data variability
    63 iris recognition system
    64 learning method
    65 least overlap-based augmentation selection protocol
    66 low error rate
    67 method
    68 model
    69 more comparisons
    70 most PAD networks
    71 network
    72 network-based iris PAD algorithms
    73 neural network-based iris PAD algorithms
    74 overlap-based augmentation selection protocol
    75 overlapping effect
    76 pad performance
    77 pads
    78 performance
    79 presentation attack detection algorithms
    80 presentation attacks
    81 protocol
    82 rate
    83 recognition system
    84 results
    85 rotation
    86 score-level combination
    87 scratch network
    88 selection protocol
    89 shift
    90 shift augmentation strategy
    91 state
    92 strategies
    93 success
    94 system
    95 technique
    96 variability
    97 vision field
    98 vulnerability
    99 schema:name The overlapping effect and fusion protocols of data augmentation techniques in iris PAD
    100 schema:pagination 8
    101 schema:productId N72e104bd78bd4298bc61a6e201c59e30
    102 Nc58af1a7665c4f6b83e264efe949b2fb
    103 schema:sameAs https://app.dimensions.ai/details/publication/pub.1143312965
    104 https://doi.org/10.1007/s00138-021-01256-9
    105 schema:sdDatePublished 2022-01-01T19:01
    106 schema:sdLicense https://scigraph.springernature.com/explorer/license/
    107 schema:sdPublisher Ndfd725e55c1e44679abccd3b2e542eb1
    108 schema:url https://doi.org/10.1007/s00138-021-01256-9
    109 sgo:license sg:explorer/license/
    110 sgo:sdDataset articles
    111 rdf:type schema:ScholarlyArticle
    112 N031de2a355bb4221bdf43b4bd55a7d36 rdf:first sg:person.07371035657.43
    113 rdf:rest N3cfbccb6885e4fdbb7f4b4b7e837b922
    114 N3cfbccb6885e4fdbb7f4b4b7e837b922 rdf:first sg:person.015632044271.26
    115 rdf:rest rdf:nil
    116 N4320566f995a49e9a9337dd909244d86 rdf:first sg:person.015017674621.27
    117 rdf:rest N723fcf938dcf4b7791cd5dd85e4cd648
    118 N723fcf938dcf4b7791cd5dd85e4cd648 rdf:first sg:person.011045327242.12
    119 rdf:rest Ncb2b8b12adf74eefb6388ee4e93ab0c9
    120 N72e104bd78bd4298bc61a6e201c59e30 schema:name doi
    121 schema:value 10.1007/s00138-021-01256-9
    122 rdf:type schema:PropertyValue
    123 N7385d04271a34d43a6766dbc9e4a6528 schema:issueNumber 1
    124 rdf:type schema:PublicationIssue
    125 Nb74230a12272428a9d5b1a8878bf94f9 schema:volumeNumber 33
    126 rdf:type schema:PublicationVolume
    127 Nc58af1a7665c4f6b83e264efe949b2fb schema:name dimensions_id
    128 schema:value pub.1143312965
    129 rdf:type schema:PropertyValue
    130 Ncb2b8b12adf74eefb6388ee4e93ab0c9 rdf:first sg:person.012172532457.60
    131 rdf:rest N031de2a355bb4221bdf43b4bd55a7d36
    132 Ndfd725e55c1e44679abccd3b2e542eb1 schema:name Springer Nature - SN SciGraph project
    133 rdf:type schema:Organization
    134 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
    135 schema:name Information and Computing Sciences
    136 rdf:type schema:DefinedTerm
    137 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
    138 schema:name Artificial Intelligence and Image Processing
    139 rdf:type schema:DefinedTerm
    140 sg:journal.1045266 schema:issn 0932-8092
    141 1432-1769
    142 schema:name Machine Vision and Applications
    143 schema:publisher Springer Nature
    144 rdf:type schema:Periodical
    145 sg:person.011045327242.12 schema:affiliation grid-institutes:grid.6546.1
    146 schema:familyName Damer
    147 schema:givenName Naser
    148 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.011045327242.12
    149 rdf:type schema:Person
    150 sg:person.012172532457.60 schema:affiliation grid-institutes:grid.6546.1
    151 schema:familyName Boutros
    152 schema:givenName Fadi
    153 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.012172532457.60
    154 rdf:type schema:Person
    155 sg:person.015017674621.27 schema:affiliation grid-institutes:grid.6546.1
    156 schema:familyName Fang
    157 schema:givenName Meiling
    158 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.015017674621.27
    159 rdf:type schema:Person
    160 sg:person.015632044271.26 schema:affiliation grid-institutes:grid.6546.1
    161 schema:familyName Kuijper
    162 schema:givenName Arjan
    163 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.015632044271.26
    164 rdf:type schema:Person
    165 sg:person.07371035657.43 schema:affiliation grid-institutes:grid.461618.c
    166 schema:familyName Kirchbuchner
    167 schema:givenName Florian
    168 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.07371035657.43
    169 rdf:type schema:Person
    170 sg:pub.10.1007/s00521-020-05342-3 schema:sameAs https://app.dimensions.ai/details/publication/pub.1131099745
    171 https://doi.org/10.1007/s00521-020-05342-3
    172 rdf:type schema:CreativeWork
    173 sg:pub.10.1007/s11042-017-4668-z schema:sameAs https://app.dimensions.ai/details/publication/pub.1085027903
    174 https://doi.org/10.1007/s11042-017-4668-z
    175 rdf:type schema:CreativeWork
    176 sg:pub.10.1186/s40537-019-0197-0 schema:sameAs https://app.dimensions.ai/details/publication/pub.1117799804
    177 https://doi.org/10.1186/s40537-019-0197-0
    178 rdf:type schema:CreativeWork
    179 grid-institutes:grid.461618.c schema:alternateName Fraunhofer Institute for Computer Graphics Research IGD, Darmstadt, Germany
    180 schema:name Fraunhofer Institute for Computer Graphics Research IGD, Darmstadt, Germany
    181 rdf:type schema:Organization
    182 grid-institutes:grid.6546.1 schema:alternateName Mathematical and Applied Visual Computing, TU Darmstadt, Darmstadt, Germany
    183 schema:name Fraunhofer Institute for Computer Graphics Research IGD, Darmstadt, Germany
    184 Mathematical and Applied Visual Computing, TU Darmstadt, Darmstadt, Germany
    185 rdf:type schema:Organization
     




    Preview window. Press ESC to close (or click here)


    ...