Deep Multiscale Convolutional Feature Learning for Weakly Supervised Localization of Chest Pathologies in X-ray Images View Full Text


Ontology type: schema:Chapter      Open Access: True


Chapter Info

DATE

2018

AUTHORS

Suman Sedai , Dwarikanath Mahapatra , Zongyuan Ge , Rajib Chakravorty , Rahil Garnavi

ABSTRACT

Localization of chest pathologies in chest X-ray images is a challenging task because of their varying sizes and appearances. We propose a novel weakly supervised method to localize chest pathologies using class aware deep multiscale feature learning. Our method leverages intermediate feature maps from CNN layers at different stages of a deep network during the training of a classification model using image level annotations of pathologies. During the training phase, a set of layer relevance weights are learned for each pathology class and the CNN is optimized to perform pathology classification by convex combination of feature maps from both shallow and deep layers using the learned weights. During the test phase, to localize the predicted pathology, the multiscale attention map is obtained by convex combination of class activation maps from each stage using the layer relevance weights learned during the training phase. We have validated our method using 112000 X-ray images and compared with the state-of-the-art localization methods. We experimentally demonstrate that the proposed weakly supervised method can improve the localization performance of small pathologies such as nodule and mass while giving comparable performance for bigger pathologies e.g., Cardiomegaly. More... »

PAGES

267-275

References to SciGraph publications

  • 2017. Accurate Pulmonary Nodule Detection in Computed Tomography Images Using Deep Convolutional Neural Networks in MEDICAL IMAGE COMPUTING AND COMPUTER-ASSISTED INTERVENTION − MICCAI 2017
  • 2017. Semi-supervised Segmentation of Optic Cup in Retinal Fundus Images Using Variational Autoencoder in MEDICAL IMAGE COMPUTING AND COMPUTER-ASSISTED INTERVENTION − MICCAI 2017
  • 2016. Self-Transfer Learning for Weakly Supervised Lesion Localization in MEDICAL IMAGE COMPUTING AND COMPUTER-ASSISTED INTERVENTION – MICCAI 2016
  • 2011. X-ray Categorization and Spatial Localization of Chest Pathologies in MEDICAL IMAGE COMPUTING AND COMPUTER-ASSISTED INTERVENTION – MICCAI 2011
  • 2015. U-Net: Convolutional Networks for Biomedical Image Segmentation in MEDICAL IMAGE COMPUTING AND COMPUTER-ASSISTED INTERVENTION – MICCAI 2015
  • Book

    TITLE

    Machine Learning in Medical Imaging

    ISBN

    978-3-030-00918-2
    978-3-030-00919-9

    Author Affiliations

    Identifiers

    URI

    http://scigraph.springernature.com/pub.10.1007/978-3-030-00919-9_31

    DOI

    http://dx.doi.org/10.1007/978-3-030-00919-9_31

    DIMENSIONS

    https://app.dimensions.ai/details/publication/pub.1107037874


    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/2008/0801", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Artificial Intelligence and Image Processing", 
            "type": "DefinedTerm"
          }, 
          {
            "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"
          }
        ], 
        "author": [
          {
            "affiliation": {
              "alternateName": "IBM Research - Australia", 
              "id": "https://www.grid.ac/institutes/grid.481553.e", 
              "name": [
                "IBM Research - Australia"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Sedai", 
            "givenName": "Suman", 
            "id": "sg:person.0754065601.77", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.0754065601.77"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "IBM Research - Australia", 
              "id": "https://www.grid.ac/institutes/grid.481553.e", 
              "name": [
                "IBM Research - Australia"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Mahapatra", 
            "givenName": "Dwarikanath", 
            "id": "sg:person.01100662063.91", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01100662063.91"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "IBM Research - Australia", 
              "id": "https://www.grid.ac/institutes/grid.481553.e", 
              "name": [
                "IBM Research - Australia"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Ge", 
            "givenName": "Zongyuan", 
            "id": "sg:person.013723142763.21", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013723142763.21"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "IBM Research - Australia", 
              "id": "https://www.grid.ac/institutes/grid.481553.e", 
              "name": [
                "IBM Research - Australia"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Chakravorty", 
            "givenName": "Rajib", 
            "id": "sg:person.01267151770.34", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01267151770.34"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "IBM Research - Australia", 
              "id": "https://www.grid.ac/institutes/grid.481553.e", 
              "name": [
                "IBM Research - Australia"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Garnavi", 
            "givenName": "Rahil", 
            "id": "sg:person.01337507642.38", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01337507642.38"
            ], 
            "type": "Person"
          }
        ], 
        "citation": [
          {
            "id": "sg:pub.10.1007/978-3-319-24574-4_28", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1017774818", 
              "https://doi.org/10.1007/978-3-319-24574-4_28"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-23626-6_25", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1024275446", 
              "https://doi.org/10.1007/978-3-642-23626-6_25"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-23626-6_25", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1024275446", 
              "https://doi.org/10.1007/978-3-642-23626-6_25"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1016/j.media.2014.01.010", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1037010598"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/tmi.2012.2230018", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1061696022"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-319-46723-8_28", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1084904229", 
              "https://doi.org/10.1007/978-3-319-46723-8_28"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-319-66179-7_64", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1091429726", 
              "https://doi.org/10.1007/978-3-319-66179-7_64"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-319-66185-8_9", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1091432167", 
              "https://doi.org/10.1007/978-3-319-66185-8_9"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/cvpr.2016.319", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1093270996"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/cvpr.2015.7298668", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1093337962"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "https://doi.org/10.1109/isbi.2017.7950704", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1094372271"
            ], 
            "type": "CreativeWork"
          }
        ], 
        "datePublished": "2018", 
        "datePublishedReg": "2018-01-01", 
        "description": "Localization of chest pathologies in chest X-ray images is a challenging task because of their varying sizes and appearances. We propose a novel weakly supervised method to localize chest pathologies using class aware deep multiscale feature learning. Our method leverages intermediate feature maps from CNN layers at different stages of a deep network during the training of a classification model using image level annotations of pathologies. During the training phase, a set of layer relevance weights are learned for each pathology class and the CNN is optimized to perform pathology classification by convex combination of feature maps from both shallow and deep layers using the learned weights. During the test phase, to localize the predicted pathology, the multiscale attention map is obtained by convex combination of class activation maps from each stage using the layer relevance weights learned during the training phase. We have validated our method using 112000 X-ray images and compared with the state-of-the-art localization methods. We experimentally demonstrate that the proposed weakly supervised method can improve the localization performance of small pathologies such as nodule and mass while giving comparable performance for bigger pathologies e.g., Cardiomegaly.", 
        "editor": [
          {
            "familyName": "Shi", 
            "givenName": "Yinghuan", 
            "type": "Person"
          }, 
          {
            "familyName": "Suk", 
            "givenName": "Heung-Il", 
            "type": "Person"
          }, 
          {
            "familyName": "Liu", 
            "givenName": "Mingxia", 
            "type": "Person"
          }
        ], 
        "genre": "chapter", 
        "id": "sg:pub.10.1007/978-3-030-00919-9_31", 
        "inLanguage": [
          "en"
        ], 
        "isAccessibleForFree": true, 
        "isPartOf": {
          "isbn": [
            "978-3-030-00918-2", 
            "978-3-030-00919-9"
          ], 
          "name": "Machine Learning in Medical Imaging", 
          "type": "Book"
        }, 
        "name": "Deep Multiscale Convolutional Feature Learning for Weakly Supervised Localization of Chest Pathologies in X-ray Images", 
        "pagination": "267-275", 
        "productId": [
          {
            "name": "doi", 
            "type": "PropertyValue", 
            "value": [
              "10.1007/978-3-030-00919-9_31"
            ]
          }, 
          {
            "name": "readcube_id", 
            "type": "PropertyValue", 
            "value": [
              "89ec94868744073c5473a69b11b74ff884bf9a8285e99b9a6972ea700995e365"
            ]
          }, 
          {
            "name": "dimensions_id", 
            "type": "PropertyValue", 
            "value": [
              "pub.1107037874"
            ]
          }
        ], 
        "publisher": {
          "location": "Cham", 
          "name": "Springer International Publishing", 
          "type": "Organisation"
        }, 
        "sameAs": [
          "https://doi.org/10.1007/978-3-030-00919-9_31", 
          "https://app.dimensions.ai/details/publication/pub.1107037874"
        ], 
        "sdDataset": "chapters", 
        "sdDatePublished": "2019-04-15T15:50", 
        "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
        "sdPublisher": {
          "name": "Springer Nature - SN SciGraph project", 
          "type": "Organization"
        }, 
        "sdSource": "s3://com-uberresearch-data-dimensions-target-20181106-alternative/cleanup/v134/2549eaecd7973599484d7c17b260dba0a4ecb94b/merge/v9/a6c9fde33151104705d4d7ff012ea9563521a3ce/jats-lookup/v90/0000000001_0000000264/records_8672_00000526.jsonl", 
        "type": "Chapter", 
        "url": "http://link.springer.com/10.1007/978-3-030-00919-9_31"
      }
    ]
     

    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.10.1007/978-3-030-00919-9_31'

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

    curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/978-3-030-00919-9_31'

    Turtle is a human-readable linked data format.

    curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/978-3-030-00919-9_31'

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

    curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/978-3-030-00919-9_31'


     

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

    138 TRIPLES      23 PREDICATES      37 URIs      20 LITERALS      8 BLANK NODES

    Subject Predicate Object
    1 sg:pub.10.1007/978-3-030-00919-9_31 schema:about anzsrc-for:08
    2 anzsrc-for:0801
    3 schema:author Nc86817b9f7574bdba453ff9454151ecd
    4 schema:citation sg:pub.10.1007/978-3-319-24574-4_28
    5 sg:pub.10.1007/978-3-319-46723-8_28
    6 sg:pub.10.1007/978-3-319-66179-7_64
    7 sg:pub.10.1007/978-3-319-66185-8_9
    8 sg:pub.10.1007/978-3-642-23626-6_25
    9 https://doi.org/10.1016/j.media.2014.01.010
    10 https://doi.org/10.1109/cvpr.2015.7298668
    11 https://doi.org/10.1109/cvpr.2016.319
    12 https://doi.org/10.1109/isbi.2017.7950704
    13 https://doi.org/10.1109/tmi.2012.2230018
    14 schema:datePublished 2018
    15 schema:datePublishedReg 2018-01-01
    16 schema:description Localization of chest pathologies in chest X-ray images is a challenging task because of their varying sizes and appearances. We propose a novel weakly supervised method to localize chest pathologies using class aware deep multiscale feature learning. Our method leverages intermediate feature maps from CNN layers at different stages of a deep network during the training of a classification model using image level annotations of pathologies. During the training phase, a set of layer relevance weights are learned for each pathology class and the CNN is optimized to perform pathology classification by convex combination of feature maps from both shallow and deep layers using the learned weights. During the test phase, to localize the predicted pathology, the multiscale attention map is obtained by convex combination of class activation maps from each stage using the layer relevance weights learned during the training phase. We have validated our method using 112000 X-ray images and compared with the state-of-the-art localization methods. We experimentally demonstrate that the proposed weakly supervised method can improve the localization performance of small pathologies such as nodule and mass while giving comparable performance for bigger pathologies e.g., Cardiomegaly.
    17 schema:editor Nb7335209ddbd4c6694e18bc0274d20f8
    18 schema:genre chapter
    19 schema:inLanguage en
    20 schema:isAccessibleForFree true
    21 schema:isPartOf N6428ffdf909e425c9096f6d948cb0bbc
    22 schema:name Deep Multiscale Convolutional Feature Learning for Weakly Supervised Localization of Chest Pathologies in X-ray Images
    23 schema:pagination 267-275
    24 schema:productId N024f55b2155d42bea5a7dcfc4d29708c
    25 N7084f7382c734caf9390aed47a88a096
    26 N75250d540c0244789629ef70de45bd5c
    27 schema:publisher Ne4423c2c7eae4edab0c416bc35264b28
    28 schema:sameAs https://app.dimensions.ai/details/publication/pub.1107037874
    29 https://doi.org/10.1007/978-3-030-00919-9_31
    30 schema:sdDatePublished 2019-04-15T15:50
    31 schema:sdLicense https://scigraph.springernature.com/explorer/license/
    32 schema:sdPublisher Nc6f1cfdf5acf425a81b8a9ed48d17dd2
    33 schema:url http://link.springer.com/10.1007/978-3-030-00919-9_31
    34 sgo:license sg:explorer/license/
    35 sgo:sdDataset chapters
    36 rdf:type schema:Chapter
    37 N024f55b2155d42bea5a7dcfc4d29708c schema:name dimensions_id
    38 schema:value pub.1107037874
    39 rdf:type schema:PropertyValue
    40 N42a28c16d75f4fd6bcb37a77531a858a rdf:first sg:person.01337507642.38
    41 rdf:rest rdf:nil
    42 N536ca9b7e1db4372a57a892fdf7930a5 rdf:first Na5b9abdf42c0440fae773406a4457672
    43 rdf:rest rdf:nil
    44 N6428ffdf909e425c9096f6d948cb0bbc schema:isbn 978-3-030-00918-2
    45 978-3-030-00919-9
    46 schema:name Machine Learning in Medical Imaging
    47 rdf:type schema:Book
    48 N7084f7382c734caf9390aed47a88a096 schema:name doi
    49 schema:value 10.1007/978-3-030-00919-9_31
    50 rdf:type schema:PropertyValue
    51 N75250d540c0244789629ef70de45bd5c schema:name readcube_id
    52 schema:value 89ec94868744073c5473a69b11b74ff884bf9a8285e99b9a6972ea700995e365
    53 rdf:type schema:PropertyValue
    54 N7a7c04a396ce411f9f40f43309578463 rdf:first sg:person.01267151770.34
    55 rdf:rest N42a28c16d75f4fd6bcb37a77531a858a
    56 N96b252e20d7246deae493cfe38b56796 rdf:first sg:person.01100662063.91
    57 rdf:rest Nf656b73971574da2b7e864fb87d344f8
    58 N9a9be3d53bba4bdb992c17f8af2e5edb rdf:first Na7d0e0b13dde406da1610ad4c998b2aa
    59 rdf:rest N536ca9b7e1db4372a57a892fdf7930a5
    60 Na5b9abdf42c0440fae773406a4457672 schema:familyName Liu
    61 schema:givenName Mingxia
    62 rdf:type schema:Person
    63 Na7d0e0b13dde406da1610ad4c998b2aa schema:familyName Suk
    64 schema:givenName Heung-Il
    65 rdf:type schema:Person
    66 Nb7335209ddbd4c6694e18bc0274d20f8 rdf:first Nc9ed39660c4743be981de0bdda26b244
    67 rdf:rest N9a9be3d53bba4bdb992c17f8af2e5edb
    68 Nc6f1cfdf5acf425a81b8a9ed48d17dd2 schema:name Springer Nature - SN SciGraph project
    69 rdf:type schema:Organization
    70 Nc86817b9f7574bdba453ff9454151ecd rdf:first sg:person.0754065601.77
    71 rdf:rest N96b252e20d7246deae493cfe38b56796
    72 Nc9ed39660c4743be981de0bdda26b244 schema:familyName Shi
    73 schema:givenName Yinghuan
    74 rdf:type schema:Person
    75 Ne4423c2c7eae4edab0c416bc35264b28 schema:location Cham
    76 schema:name Springer International Publishing
    77 rdf:type schema:Organisation
    78 Nf656b73971574da2b7e864fb87d344f8 rdf:first sg:person.013723142763.21
    79 rdf:rest N7a7c04a396ce411f9f40f43309578463
    80 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
    81 schema:name Information and Computing Sciences
    82 rdf:type schema:DefinedTerm
    83 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
    84 schema:name Artificial Intelligence and Image Processing
    85 rdf:type schema:DefinedTerm
    86 sg:person.01100662063.91 schema:affiliation https://www.grid.ac/institutes/grid.481553.e
    87 schema:familyName Mahapatra
    88 schema:givenName Dwarikanath
    89 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01100662063.91
    90 rdf:type schema:Person
    91 sg:person.01267151770.34 schema:affiliation https://www.grid.ac/institutes/grid.481553.e
    92 schema:familyName Chakravorty
    93 schema:givenName Rajib
    94 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01267151770.34
    95 rdf:type schema:Person
    96 sg:person.01337507642.38 schema:affiliation https://www.grid.ac/institutes/grid.481553.e
    97 schema:familyName Garnavi
    98 schema:givenName Rahil
    99 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01337507642.38
    100 rdf:type schema:Person
    101 sg:person.013723142763.21 schema:affiliation https://www.grid.ac/institutes/grid.481553.e
    102 schema:familyName Ge
    103 schema:givenName Zongyuan
    104 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013723142763.21
    105 rdf:type schema:Person
    106 sg:person.0754065601.77 schema:affiliation https://www.grid.ac/institutes/grid.481553.e
    107 schema:familyName Sedai
    108 schema:givenName Suman
    109 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.0754065601.77
    110 rdf:type schema:Person
    111 sg:pub.10.1007/978-3-319-24574-4_28 schema:sameAs https://app.dimensions.ai/details/publication/pub.1017774818
    112 https://doi.org/10.1007/978-3-319-24574-4_28
    113 rdf:type schema:CreativeWork
    114 sg:pub.10.1007/978-3-319-46723-8_28 schema:sameAs https://app.dimensions.ai/details/publication/pub.1084904229
    115 https://doi.org/10.1007/978-3-319-46723-8_28
    116 rdf:type schema:CreativeWork
    117 sg:pub.10.1007/978-3-319-66179-7_64 schema:sameAs https://app.dimensions.ai/details/publication/pub.1091429726
    118 https://doi.org/10.1007/978-3-319-66179-7_64
    119 rdf:type schema:CreativeWork
    120 sg:pub.10.1007/978-3-319-66185-8_9 schema:sameAs https://app.dimensions.ai/details/publication/pub.1091432167
    121 https://doi.org/10.1007/978-3-319-66185-8_9
    122 rdf:type schema:CreativeWork
    123 sg:pub.10.1007/978-3-642-23626-6_25 schema:sameAs https://app.dimensions.ai/details/publication/pub.1024275446
    124 https://doi.org/10.1007/978-3-642-23626-6_25
    125 rdf:type schema:CreativeWork
    126 https://doi.org/10.1016/j.media.2014.01.010 schema:sameAs https://app.dimensions.ai/details/publication/pub.1037010598
    127 rdf:type schema:CreativeWork
    128 https://doi.org/10.1109/cvpr.2015.7298668 schema:sameAs https://app.dimensions.ai/details/publication/pub.1093337962
    129 rdf:type schema:CreativeWork
    130 https://doi.org/10.1109/cvpr.2016.319 schema:sameAs https://app.dimensions.ai/details/publication/pub.1093270996
    131 rdf:type schema:CreativeWork
    132 https://doi.org/10.1109/isbi.2017.7950704 schema:sameAs https://app.dimensions.ai/details/publication/pub.1094372271
    133 rdf:type schema:CreativeWork
    134 https://doi.org/10.1109/tmi.2012.2230018 schema:sameAs https://app.dimensions.ai/details/publication/pub.1061696022
    135 rdf:type schema:CreativeWork
    136 https://www.grid.ac/institutes/grid.481553.e schema:alternateName IBM Research - Australia
    137 schema:name IBM Research - Australia
    138 rdf:type schema:Organization
     




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


    ...