An Artificial Agent for Anatomical Landmark Detection in Medical Images View Full Text


Ontology type: schema:Chapter      Open Access: True


Chapter Info

DATE

2016-10-02

AUTHORS

Florin C. Ghesu , Bogdan Georgescu , Tommaso Mansi , Dominik Neumann , Joachim Hornegger , Dorin Comaniciu

ABSTRACT

Fast and robust detection of anatomical structures or pathologies represents a fundamental task in medical image analysis. Most of the current solutions are however suboptimal and unconstrained by learning an appearance model and exhaustively scanning the space of parameters to detect a specific anatomical structure. In addition, typical feature computation or estimation of meta-parameters related to the appearance model or the search strategy, is based on local criteria or predefined approximation schemes. We propose a new learning method following a fundamentally different paradigm by simultaneously modeling both the object appearance and the parameter search strategy as a unified behavioral task for an artificial agent. The method combines the advantages of behavior learning achieved through reinforcement learning with effective hierarchical feature extraction achieved through deep learning. We show that given only a sequence of annotated images, the agent can automatically and strategically learn optimal paths that converge to the sought anatomical landmark location as opposed to exhaustively scanning the entire solution space. The method significantly outperforms state-of-the-art machine learning and deep learning approaches both in terms of accuracy and speed on 2D magnetic resonance images, 2D ultrasound and 3D CT images, achieving average detection errors of 1-2 pixels, while also recognizing the absence of an object from the image. More... »

PAGES

229-237

Identifiers

URI

http://scigraph.springernature.com/pub.10.1007/978-3-319-46726-9_27

DOI

http://dx.doi.org/10.1007/978-3-319-46726-9_27

DIMENSIONS

https://app.dimensions.ai/details/publication/pub.1084911791


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": "Pattern Recognition Lab, Friedrich-Alexander-Universit\u00e4t, Erlangen, Germany", 
          "id": "http://www.grid.ac/institutes/grid.5330.5", 
          "name": [
            "Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA", 
            "Pattern Recognition Lab, Friedrich-Alexander-Universit\u00e4t, Erlangen, Germany"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Ghesu", 
        "givenName": "Florin C.", 
        "id": "sg:person.012717301041.17", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.012717301041.17"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA", 
          "id": "http://www.grid.ac/institutes/grid.415886.6", 
          "name": [
            "Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Georgescu", 
        "givenName": "Bogdan", 
        "id": "sg:person.0703547214.37", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.0703547214.37"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA", 
          "id": "http://www.grid.ac/institutes/grid.415886.6", 
          "name": [
            "Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Mansi", 
        "givenName": "Tommaso", 
        "id": "sg:person.01217474726.73", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01217474726.73"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA", 
          "id": "http://www.grid.ac/institutes/grid.415886.6", 
          "name": [
            "Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Neumann", 
        "givenName": "Dominik", 
        "id": "sg:person.01054566020.28", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01054566020.28"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Pattern Recognition Lab, Friedrich-Alexander-Universit\u00e4t, Erlangen, Germany", 
          "id": "http://www.grid.ac/institutes/grid.5330.5", 
          "name": [
            "Pattern Recognition Lab, Friedrich-Alexander-Universit\u00e4t, Erlangen, Germany"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Hornegger", 
        "givenName": "Joachim", 
        "id": "sg:person.01322323610.92", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01322323610.92"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA", 
          "id": "http://www.grid.ac/institutes/grid.415886.6", 
          "name": [
            "Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Comaniciu", 
        "givenName": "Dorin", 
        "id": "sg:person.01066111014.77", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01066111014.77"
        ], 
        "type": "Person"
      }
    ], 
    "datePublished": "2016-10-02", 
    "datePublishedReg": "2016-10-02", 
    "description": "Fast and robust detection of anatomical structures or pathologies represents a fundamental task in medical image analysis. Most of the current solutions are however suboptimal and unconstrained by learning an appearance model and exhaustively scanning the space of parameters to detect a specific anatomical structure. In addition, typical feature computation or estimation of meta-parameters related to the appearance model or the search strategy, is based on local criteria or predefined approximation schemes. We propose a new learning method following a fundamentally different paradigm by simultaneously modeling both the object appearance and the parameter search strategy as a unified behavioral task for an artificial agent. The method combines the advantages of behavior learning achieved through reinforcement learning with effective hierarchical feature extraction achieved through deep learning. We show that given only a sequence of annotated images, the agent can automatically and strategically learn optimal paths that converge to the sought anatomical landmark location as opposed to exhaustively scanning the entire solution space. The method significantly outperforms state-of-the-art machine learning and deep learning approaches both in terms of accuracy and speed on 2D magnetic resonance images, 2D ultrasound and 3D CT images, achieving average detection errors of 1-2 pixels, while also recognizing the absence of an object from the image.", 
    "editor": [
      {
        "familyName": "Ourselin", 
        "givenName": "Sebastien", 
        "type": "Person"
      }, 
      {
        "familyName": "Joskowicz", 
        "givenName": "Leo", 
        "type": "Person"
      }, 
      {
        "familyName": "Sabuncu", 
        "givenName": "Mert R.", 
        "type": "Person"
      }, 
      {
        "familyName": "Unal", 
        "givenName": "Gozde", 
        "type": "Person"
      }, 
      {
        "familyName": "Wells", 
        "givenName": "William", 
        "type": "Person"
      }
    ], 
    "genre": "chapter", 
    "id": "sg:pub.10.1007/978-3-319-46726-9_27", 
    "inLanguage": "en", 
    "isAccessibleForFree": true, 
    "isPartOf": {
      "isbn": [
        "978-3-319-46725-2", 
        "978-3-319-46726-9"
      ], 
      "name": "Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016", 
      "type": "Book"
    }, 
    "keywords": [
      "deep learning", 
      "artificial agents", 
      "appearance model", 
      "medical image analysis", 
      "anatomical landmark detection", 
      "hierarchical feature extraction", 
      "art machine learning", 
      "average detection error", 
      "parameter search strategy", 
      "terms of accuracy", 
      "entire solution space", 
      "new learning method", 
      "feature computation", 
      "search strategy", 
      "behavior learning", 
      "machine learning", 
      "landmark detection", 
      "medical images", 
      "feature extraction", 
      "learning method", 
      "current solutions", 
      "object appearance", 
      "fundamental task", 
      "optimal path", 
      "landmark locations", 
      "anatomical landmark locations", 
      "solution space", 
      "robust detection", 
      "detection error", 
      "image analysis", 
      "learning", 
      "images", 
      "anatomical structures", 
      "specific anatomical structures", 
      "magnetic resonance images", 
      "different paradigms", 
      "task", 
      "approximation scheme", 
      "CT images", 
      "resonance images", 
      "local criteria", 
      "pixels", 
      "computation", 
      "detection", 
      "objects", 
      "scheme", 
      "space", 
      "paradigm", 
      "accuracy", 
      "method", 
      "space of parameters", 
      "model", 
      "path", 
      "extraction", 
      "error", 
      "advantages", 
      "speed", 
      "estimation", 
      "strategies", 
      "solution", 
      "location", 
      "agents", 
      "terms", 
      "sequence", 
      "structure", 
      "state", 
      "parameters", 
      "criteria", 
      "reinforcement", 
      "analysis", 
      "addition", 
      "appearance", 
      "behavioral tasks", 
      "ultrasound", 
      "pathology", 
      "absence", 
      "typical feature computation", 
      "unified behavioral task", 
      "effective hierarchical feature extraction"
    ], 
    "name": "An Artificial Agent for Anatomical Landmark Detection in Medical Images", 
    "pagination": "229-237", 
    "productId": [
      {
        "name": "dimensions_id", 
        "type": "PropertyValue", 
        "value": [
          "pub.1084911791"
        ]
      }, 
      {
        "name": "doi", 
        "type": "PropertyValue", 
        "value": [
          "10.1007/978-3-319-46726-9_27"
        ]
      }
    ], 
    "publisher": {
      "name": "Springer Nature", 
      "type": "Organisation"
    }, 
    "sameAs": [
      "https://doi.org/10.1007/978-3-319-46726-9_27", 
      "https://app.dimensions.ai/details/publication/pub.1084911791"
    ], 
    "sdDataset": "chapters", 
    "sdDatePublished": "2022-01-01T19:28", 
    "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/chapter/chapter_80.jsonl", 
    "type": "Chapter", 
    "url": "https://doi.org/10.1007/978-3-319-46726-9_27"
  }
]
 

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.10.1007/978-3-319-46726-9_27'

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

curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/978-3-319-46726-9_27'

Turtle is a human-readable linked data format.

curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/978-3-319-46726-9_27'

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

curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/978-3-319-46726-9_27'


 

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

198 TRIPLES      23 PREDICATES      104 URIs      97 LITERALS      7 BLANK NODES

Subject Predicate Object
1 sg:pub.10.1007/978-3-319-46726-9_27 schema:about anzsrc-for:08
2 anzsrc-for:0801
3 schema:author N0bb294d74ee34d79bb18c1a4b26adbeb
4 schema:datePublished 2016-10-02
5 schema:datePublishedReg 2016-10-02
6 schema:description Fast and robust detection of anatomical structures or pathologies represents a fundamental task in medical image analysis. Most of the current solutions are however suboptimal and unconstrained by learning an appearance model and exhaustively scanning the space of parameters to detect a specific anatomical structure. In addition, typical feature computation or estimation of meta-parameters related to the appearance model or the search strategy, is based on local criteria or predefined approximation schemes. We propose a new learning method following a fundamentally different paradigm by simultaneously modeling both the object appearance and the parameter search strategy as a unified behavioral task for an artificial agent. The method combines the advantages of behavior learning achieved through reinforcement learning with effective hierarchical feature extraction achieved through deep learning. We show that given only a sequence of annotated images, the agent can automatically and strategically learn optimal paths that converge to the sought anatomical landmark location as opposed to exhaustively scanning the entire solution space. The method significantly outperforms state-of-the-art machine learning and deep learning approaches both in terms of accuracy and speed on 2D magnetic resonance images, 2D ultrasound and 3D CT images, achieving average detection errors of 1-2 pixels, while also recognizing the absence of an object from the image.
7 schema:editor N476e61a55b834a13b58f3d184843ed5c
8 schema:genre chapter
9 schema:inLanguage en
10 schema:isAccessibleForFree true
11 schema:isPartOf N9214a4c0761f4318b06eba93278d63b6
12 schema:keywords CT images
13 absence
14 accuracy
15 addition
16 advantages
17 agents
18 analysis
19 anatomical landmark detection
20 anatomical landmark locations
21 anatomical structures
22 appearance
23 appearance model
24 approximation scheme
25 art machine learning
26 artificial agents
27 average detection error
28 behavior learning
29 behavioral tasks
30 computation
31 criteria
32 current solutions
33 deep learning
34 detection
35 detection error
36 different paradigms
37 effective hierarchical feature extraction
38 entire solution space
39 error
40 estimation
41 extraction
42 feature computation
43 feature extraction
44 fundamental task
45 hierarchical feature extraction
46 image analysis
47 images
48 landmark detection
49 landmark locations
50 learning
51 learning method
52 local criteria
53 location
54 machine learning
55 magnetic resonance images
56 medical image analysis
57 medical images
58 method
59 model
60 new learning method
61 object appearance
62 objects
63 optimal path
64 paradigm
65 parameter search strategy
66 parameters
67 path
68 pathology
69 pixels
70 reinforcement
71 resonance images
72 robust detection
73 scheme
74 search strategy
75 sequence
76 solution
77 solution space
78 space
79 space of parameters
80 specific anatomical structures
81 speed
82 state
83 strategies
84 structure
85 task
86 terms
87 terms of accuracy
88 typical feature computation
89 ultrasound
90 unified behavioral task
91 schema:name An Artificial Agent for Anatomical Landmark Detection in Medical Images
92 schema:pagination 229-237
93 schema:productId Na7b0234194124071b5a2b8be9d83e3b0
94 Nfe52f9231e5b4bbdaa94f5c659560434
95 schema:publisher N200b629a5a7b4c7195132bf4227673ea
96 schema:sameAs https://app.dimensions.ai/details/publication/pub.1084911791
97 https://doi.org/10.1007/978-3-319-46726-9_27
98 schema:sdDatePublished 2022-01-01T19:28
99 schema:sdLicense https://scigraph.springernature.com/explorer/license/
100 schema:sdPublisher N5be438ee6992463a97791af86df44f58
101 schema:url https://doi.org/10.1007/978-3-319-46726-9_27
102 sgo:license sg:explorer/license/
103 sgo:sdDataset chapters
104 rdf:type schema:Chapter
105 N0a0bde2392fc4ba5bf77124eb420fb92 schema:familyName Unal
106 schema:givenName Gozde
107 rdf:type schema:Person
108 N0bb294d74ee34d79bb18c1a4b26adbeb rdf:first sg:person.012717301041.17
109 rdf:rest Neaa43430033541adbec096f4d554f0c3
110 N11d3255fd10546a4ad44d62fd1f40423 schema:familyName Joskowicz
111 schema:givenName Leo
112 rdf:type schema:Person
113 N17d5462bc3924b90bf30fc6196833bd8 schema:familyName Ourselin
114 schema:givenName Sebastien
115 rdf:type schema:Person
116 N1fc780085fce475f90688f3f6942c8ef rdf:first sg:person.01217474726.73
117 rdf:rest Nb4d8c49953c640cebc8562cc5a663e02
118 N200b629a5a7b4c7195132bf4227673ea schema:name Springer Nature
119 rdf:type schema:Organisation
120 N235a8f1949854a0d8597bd4e3d2c3289 rdf:first N0a0bde2392fc4ba5bf77124eb420fb92
121 rdf:rest Nad12ce176d1644c6b7409ee1616d9f1d
122 N3d285c49d47644dc8a7bf219b965b872 rdf:first sg:person.01066111014.77
123 rdf:rest rdf:nil
124 N476e61a55b834a13b58f3d184843ed5c rdf:first N17d5462bc3924b90bf30fc6196833bd8
125 rdf:rest N940990247f504bbb8fbb4eceaae925e6
126 N51679e8ab4b14aa99fce58ac43728b4c rdf:first N878e6e784ca44d0fa3df48e82c32b1ec
127 rdf:rest N235a8f1949854a0d8597bd4e3d2c3289
128 N5be438ee6992463a97791af86df44f58 schema:name Springer Nature - SN SciGraph project
129 rdf:type schema:Organization
130 N878e6e784ca44d0fa3df48e82c32b1ec schema:familyName Sabuncu
131 schema:givenName Mert R.
132 rdf:type schema:Person
133 N9214a4c0761f4318b06eba93278d63b6 schema:isbn 978-3-319-46725-2
134 978-3-319-46726-9
135 schema:name Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016
136 rdf:type schema:Book
137 N940990247f504bbb8fbb4eceaae925e6 rdf:first N11d3255fd10546a4ad44d62fd1f40423
138 rdf:rest N51679e8ab4b14aa99fce58ac43728b4c
139 Na7b0234194124071b5a2b8be9d83e3b0 schema:name doi
140 schema:value 10.1007/978-3-319-46726-9_27
141 rdf:type schema:PropertyValue
142 Nab704ab4a0f347fcb74e62f6a022c5f4 schema:familyName Wells
143 schema:givenName William
144 rdf:type schema:Person
145 Nad12ce176d1644c6b7409ee1616d9f1d rdf:first Nab704ab4a0f347fcb74e62f6a022c5f4
146 rdf:rest rdf:nil
147 Nb4d8c49953c640cebc8562cc5a663e02 rdf:first sg:person.01054566020.28
148 rdf:rest Ncbad0a3bf1d14085902cfb37fe82132b
149 Ncbad0a3bf1d14085902cfb37fe82132b rdf:first sg:person.01322323610.92
150 rdf:rest N3d285c49d47644dc8a7bf219b965b872
151 Neaa43430033541adbec096f4d554f0c3 rdf:first sg:person.0703547214.37
152 rdf:rest N1fc780085fce475f90688f3f6942c8ef
153 Nfe52f9231e5b4bbdaa94f5c659560434 schema:name dimensions_id
154 schema:value pub.1084911791
155 rdf:type schema:PropertyValue
156 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
157 schema:name Information and Computing Sciences
158 rdf:type schema:DefinedTerm
159 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
160 schema:name Artificial Intelligence and Image Processing
161 rdf:type schema:DefinedTerm
162 sg:person.01054566020.28 schema:affiliation grid-institutes:grid.415886.6
163 schema:familyName Neumann
164 schema:givenName Dominik
165 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01054566020.28
166 rdf:type schema:Person
167 sg:person.01066111014.77 schema:affiliation grid-institutes:grid.415886.6
168 schema:familyName Comaniciu
169 schema:givenName Dorin
170 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01066111014.77
171 rdf:type schema:Person
172 sg:person.01217474726.73 schema:affiliation grid-institutes:grid.415886.6
173 schema:familyName Mansi
174 schema:givenName Tommaso
175 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01217474726.73
176 rdf:type schema:Person
177 sg:person.012717301041.17 schema:affiliation grid-institutes:grid.5330.5
178 schema:familyName Ghesu
179 schema:givenName Florin C.
180 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.012717301041.17
181 rdf:type schema:Person
182 sg:person.01322323610.92 schema:affiliation grid-institutes:grid.5330.5
183 schema:familyName Hornegger
184 schema:givenName Joachim
185 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.01322323610.92
186 rdf:type schema:Person
187 sg:person.0703547214.37 schema:affiliation grid-institutes:grid.415886.6
188 schema:familyName Georgescu
189 schema:givenName Bogdan
190 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.0703547214.37
191 rdf:type schema:Person
192 grid-institutes:grid.415886.6 schema:alternateName Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA
193 schema:name Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA
194 rdf:type schema:Organization
195 grid-institutes:grid.5330.5 schema:alternateName Pattern Recognition Lab, Friedrich-Alexander-Universität, Erlangen, Germany
196 schema:name Medical Imaging Technologies, Siemens Healthineers, Princeton, New Jersey, USA
197 Pattern Recognition Lab, Friedrich-Alexander-Universität, Erlangen, Germany
198 rdf:type schema:Organization
 




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


...