Deep Convolutional Neural Networks for breast cancer screening.
Identifieur interne : 000442 ( Main/Exploration ); précédent : 000441; suivant : 000443Deep Convolutional Neural Networks for breast cancer screening.
Auteurs : Hiba Chougrad [Maroc] ; Hamid Zouaki [Maroc] ; Omar Alheyane [Maroc]Source :
- Computer methods and programs in biomedicine [ 1872-7565 ] ; 2018.
Descripteurs français
- KwdFr :
- Apprentissage machine (MeSH), Bases de données factuelles (MeSH), Diagnostic assisté par ordinateur (MeSH), Dépistage précoce du cancer (MeSH), Femelle (MeSH), Humains (MeSH), Mammographie (méthodes), Reproductibilité des résultats (MeSH), Tumeurs du sein (anatomopathologie), Tumeurs du sein (classification), Tumeurs du sein (imagerie diagnostique).
- MESH :
- anatomopathologie : Tumeurs du sein.
- imagerie diagnostique : Tumeurs du sein.
- méthodes : Mammographie.
- classification : Apprentissage machine, Bases de données factuelles, Diagnostic assisté par ordinateur, Dépistage précoce du cancer, Femelle, Humains, Reproductibilité des résultats, Tumeurs du sein.
English descriptors
- KwdEn :
- Breast Neoplasms (classification), Breast Neoplasms (diagnostic imaging), Breast Neoplasms (pathology), Databases, Factual (MeSH), Diagnosis, Computer-Assisted (MeSH), Early Detection of Cancer (MeSH), Female (MeSH), Humans (MeSH), Machine Learning (MeSH), Mammography (methods), Neural Networks, Computer (MeSH), Reproducibility of Results (MeSH).
- MESH :
- classification : Breast Neoplasms.
- diagnostic imaging : Breast Neoplasms.
- methods : Mammography.
- pathology : Breast Neoplasms.
- Databases, Factual, Diagnosis, Computer-Assisted, Early Detection of Cancer, Female, Humans, Machine Learning, Neural Networks, Computer, Reproducibility of Results.
Abstract
BACKGROUND AND OBJECTIVE
Radiologists often have a hard time classifying mammography mass lesions which leads to unnecessary breast biopsies to remove suspicions and this ends up adding exorbitant expenses to an already burdened patient and health care system.
METHODS
In this paper we developed a Computer-aided Diagnosis (CAD) system based on deep Convolutional Neural Networks (CNN) that aims to help the radiologist classify mammography mass lesions. Deep learning usually requires large datasets to train networks of a certain depth from scratch. Transfer learning is an effective method to deal with relatively small datasets as in the case of medical images, although it can be tricky as we can easily start overfitting.
RESULTS
In this work, we explore the importance of transfer learning and we experimentally determine the best fine-tuning strategy to adopt when training a CNN model. We were able to successfully fine-tune some of the recent, most powerful CNNs and achieved better results compared to other state-of-the-art methods which classified the same public datasets. For instance we achieved 97.35% accuracy and 0.98 AUC on the DDSM database, 95.50% accuracy and 0.97 AUC on the INbreast database and 96.67% accuracy and 0.96 AUC on the BCDR database. Furthermore, after pre-processing and normalizing all the extracted Regions of Interest (ROIs) from the full mammograms, we merged all the datasets to build one large set of images and used it to fine-tune our CNNs. The CNN model which achieved the best results, a 98.94% accuracy, was used as a baseline to build the Breast Cancer Screening Framework. To evaluate the proposed CAD system and its efficiency to classify new images, we tested it on an independent database (MIAS) and got 98.23% accuracy and 0.99 AUC.
CONCLUSION
The results obtained demonstrate that the proposed framework is performant and can indeed be used to predict if the mass lesions are benign or malignant.
DOI: 10.1016/j.cmpb.2018.01.011
PubMed: 29477427
Affiliations:
Links toward previous steps (curation, corpus...)
- to stream PubMed, to step Corpus: 000369
- to stream PubMed, to step Curation: 000368
- to stream PubMed, to step Checkpoint: 000367
- to stream Main, to step Merge: 000442
- to stream Main, to step Curation: 000442
Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en">Deep Convolutional Neural Networks for breast cancer screening.</title>
<author><name sortKey="Chougrad, Hiba" sort="Chougrad, Hiba" uniqKey="Chougrad H" first="Hiba" last="Chougrad">Hiba Chougrad</name>
<affiliation wicri:level="1"><nlm:affiliation>Laboratory of Computer Science and Mathematics and their Applications (LIMA), Faculty of science, University Chouaib Doukkali, El Jadida 24000, Morocco. Electronic address: chougrad.h@ucd.ac.ma.</nlm:affiliation>
<country xml:lang="fr">Maroc</country>
<wicri:regionArea>Laboratory of Computer Science and Mathematics and their Applications (LIMA), Faculty of science, University Chouaib Doukkali, El Jadida 24000</wicri:regionArea>
<wicri:noRegion>El Jadida 24000</wicri:noRegion>
</affiliation>
</author>
<author><name sortKey="Zouaki, Hamid" sort="Zouaki, Hamid" uniqKey="Zouaki H" first="Hamid" last="Zouaki">Hamid Zouaki</name>
<affiliation wicri:level="1"><nlm:affiliation>Laboratory of Computer Science and Mathematics and their Applications (LIMA), Faculty of science, University Chouaib Doukkali, El Jadida 24000, Morocco.</nlm:affiliation>
<country xml:lang="fr">Maroc</country>
<wicri:regionArea>Laboratory of Computer Science and Mathematics and their Applications (LIMA), Faculty of science, University Chouaib Doukkali, El Jadida 24000</wicri:regionArea>
<wicri:noRegion>El Jadida 24000</wicri:noRegion>
</affiliation>
</author>
<author><name sortKey="Alheyane, Omar" sort="Alheyane, Omar" uniqKey="Alheyane O" first="Omar" last="Alheyane">Omar Alheyane</name>
<affiliation wicri:level="1"><nlm:affiliation>Laboratory of Fundamental Mathematics (LMF), Faculty of science, University Chouaib Doukkali, El Jadida 24000, Morocco.</nlm:affiliation>
<country xml:lang="fr">Maroc</country>
<wicri:regionArea>Laboratory of Fundamental Mathematics (LMF), Faculty of science, University Chouaib Doukkali, El Jadida 24000</wicri:regionArea>
<wicri:noRegion>El Jadida 24000</wicri:noRegion>
</affiliation>
</author>
</titleStmt>
<publicationStmt><idno type="wicri:source">PubMed</idno>
<date when="2018">2018</date>
<idno type="RBID">pubmed:29477427</idno>
<idno type="pmid">29477427</idno>
<idno type="doi">10.1016/j.cmpb.2018.01.011</idno>
<idno type="wicri:Area/PubMed/Corpus">000369</idno>
<idno type="wicri:explorRef" wicri:stream="PubMed" wicri:step="Corpus" wicri:corpus="PubMed">000369</idno>
<idno type="wicri:Area/PubMed/Curation">000368</idno>
<idno type="wicri:explorRef" wicri:stream="PubMed" wicri:step="Curation">000368</idno>
<idno type="wicri:Area/PubMed/Checkpoint">000367</idno>
<idno type="wicri:explorRef" wicri:stream="Checkpoint" wicri:step="PubMed">000367</idno>
<idno type="wicri:Area/Main/Merge">000442</idno>
<idno type="wicri:Area/Main/Curation">000442</idno>
<idno type="wicri:Area/Main/Exploration">000442</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title xml:lang="en">Deep Convolutional Neural Networks for breast cancer screening.</title>
<author><name sortKey="Chougrad, Hiba" sort="Chougrad, Hiba" uniqKey="Chougrad H" first="Hiba" last="Chougrad">Hiba Chougrad</name>
<affiliation wicri:level="1"><nlm:affiliation>Laboratory of Computer Science and Mathematics and their Applications (LIMA), Faculty of science, University Chouaib Doukkali, El Jadida 24000, Morocco. Electronic address: chougrad.h@ucd.ac.ma.</nlm:affiliation>
<country xml:lang="fr">Maroc</country>
<wicri:regionArea>Laboratory of Computer Science and Mathematics and their Applications (LIMA), Faculty of science, University Chouaib Doukkali, El Jadida 24000</wicri:regionArea>
<wicri:noRegion>El Jadida 24000</wicri:noRegion>
</affiliation>
</author>
<author><name sortKey="Zouaki, Hamid" sort="Zouaki, Hamid" uniqKey="Zouaki H" first="Hamid" last="Zouaki">Hamid Zouaki</name>
<affiliation wicri:level="1"><nlm:affiliation>Laboratory of Computer Science and Mathematics and their Applications (LIMA), Faculty of science, University Chouaib Doukkali, El Jadida 24000, Morocco.</nlm:affiliation>
<country xml:lang="fr">Maroc</country>
<wicri:regionArea>Laboratory of Computer Science and Mathematics and their Applications (LIMA), Faculty of science, University Chouaib Doukkali, El Jadida 24000</wicri:regionArea>
<wicri:noRegion>El Jadida 24000</wicri:noRegion>
</affiliation>
</author>
<author><name sortKey="Alheyane, Omar" sort="Alheyane, Omar" uniqKey="Alheyane O" first="Omar" last="Alheyane">Omar Alheyane</name>
<affiliation wicri:level="1"><nlm:affiliation>Laboratory of Fundamental Mathematics (LMF), Faculty of science, University Chouaib Doukkali, El Jadida 24000, Morocco.</nlm:affiliation>
<country xml:lang="fr">Maroc</country>
<wicri:regionArea>Laboratory of Fundamental Mathematics (LMF), Faculty of science, University Chouaib Doukkali, El Jadida 24000</wicri:regionArea>
<wicri:noRegion>El Jadida 24000</wicri:noRegion>
</affiliation>
</author>
</analytic>
<series><title level="j">Computer methods and programs in biomedicine</title>
<idno type="eISSN">1872-7565</idno>
<imprint><date when="2018" type="published">2018</date>
</imprint>
</series>
</biblStruct>
</sourceDesc>
</fileDesc>
<profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>Breast Neoplasms (classification)</term>
<term>Breast Neoplasms (diagnostic imaging)</term>
<term>Breast Neoplasms (pathology)</term>
<term>Databases, Factual (MeSH)</term>
<term>Diagnosis, Computer-Assisted (MeSH)</term>
<term>Early Detection of Cancer (MeSH)</term>
<term>Female (MeSH)</term>
<term>Humans (MeSH)</term>
<term>Machine Learning (MeSH)</term>
<term>Mammography (methods)</term>
<term>Neural Networks, Computer (MeSH)</term>
<term>Reproducibility of Results (MeSH)</term>
</keywords>
<keywords scheme="KwdFr" xml:lang="fr"><term>Apprentissage machine (MeSH)</term>
<term>Bases de données factuelles (MeSH)</term>
<term>Diagnostic assisté par ordinateur (MeSH)</term>
<term>Dépistage précoce du cancer (MeSH)</term>
<term>Femelle (MeSH)</term>
<term>Humains (MeSH)</term>
<term>Mammographie (méthodes)</term>
<term>Reproductibilité des résultats (MeSH)</term>
<term>Tumeurs du sein (anatomopathologie)</term>
<term>Tumeurs du sein (classification)</term>
<term>Tumeurs du sein (imagerie diagnostique)</term>
</keywords>
<keywords scheme="MESH" qualifier="anatomopathologie" xml:lang="fr"><term>Tumeurs du sein</term>
</keywords>
<keywords scheme="MESH" qualifier="classification" xml:lang="en"><term>Breast Neoplasms</term>
</keywords>
<keywords scheme="MESH" qualifier="diagnostic imaging" xml:lang="en"><term>Breast Neoplasms</term>
</keywords>
<keywords scheme="MESH" qualifier="imagerie diagnostique" xml:lang="fr"><term>Tumeurs du sein</term>
</keywords>
<keywords scheme="MESH" qualifier="methods" xml:lang="en"><term>Mammography</term>
</keywords>
<keywords scheme="MESH" qualifier="méthodes" xml:lang="fr"><term>Mammographie</term>
</keywords>
<keywords scheme="MESH" qualifier="pathology" xml:lang="en"><term>Breast Neoplasms</term>
</keywords>
<keywords scheme="MESH" xml:lang="en"><term>Databases, Factual</term>
<term>Diagnosis, Computer-Assisted</term>
<term>Early Detection of Cancer</term>
<term>Female</term>
<term>Humans</term>
<term>Machine Learning</term>
<term>Neural Networks, Computer</term>
<term>Reproducibility of Results</term>
</keywords>
<keywords scheme="MESH" qualifier="classification" xml:lang="fr"><term>Apprentissage machine</term>
<term>Bases de données factuelles</term>
<term>Diagnostic assisté par ordinateur</term>
<term>Dépistage précoce du cancer</term>
<term>Femelle</term>
<term>Humains</term>
<term>Reproductibilité des résultats</term>
<term>Tumeurs du sein</term>
</keywords>
</textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en"><p><b>BACKGROUND AND OBJECTIVE</b>
</p>
<p>Radiologists often have a hard time classifying mammography mass lesions which leads to unnecessary breast biopsies to remove suspicions and this ends up adding exorbitant expenses to an already burdened patient and health care system.</p>
</div>
<div type="abstract" xml:lang="en"><p><b>METHODS</b>
</p>
<p>In this paper we developed a Computer-aided Diagnosis (CAD) system based on deep Convolutional Neural Networks (CNN) that aims to help the radiologist classify mammography mass lesions. Deep learning usually requires large datasets to train networks of a certain depth from scratch. Transfer learning is an effective method to deal with relatively small datasets as in the case of medical images, although it can be tricky as we can easily start overfitting.</p>
</div>
<div type="abstract" xml:lang="en"><p><b>RESULTS</b>
</p>
<p>In this work, we explore the importance of transfer learning and we experimentally determine the best fine-tuning strategy to adopt when training a CNN model. We were able to successfully fine-tune some of the recent, most powerful CNNs and achieved better results compared to other state-of-the-art methods which classified the same public datasets. For instance we achieved 97.35% accuracy and 0.98 AUC on the DDSM database, 95.50% accuracy and 0.97 AUC on the INbreast database and 96.67% accuracy and 0.96 AUC on the BCDR database. Furthermore, after pre-processing and normalizing all the extracted Regions of Interest (ROIs) from the full mammograms, we merged all the datasets to build one large set of images and used it to fine-tune our CNNs. The CNN model which achieved the best results, a 98.94% accuracy, was used as a baseline to build the Breast Cancer Screening Framework. To evaluate the proposed CAD system and its efficiency to classify new images, we tested it on an independent database (MIAS) and got 98.23% accuracy and 0.99 AUC.</p>
</div>
<div type="abstract" xml:lang="en"><p><b>CONCLUSION</b>
</p>
<p>The results obtained demonstrate that the proposed framework is performant and can indeed be used to predict if the mass lesions are benign or malignant.</p>
</div>
</front>
</TEI>
<affiliations><list><country><li>Maroc</li>
</country>
</list>
<tree><country name="Maroc"><noRegion><name sortKey="Chougrad, Hiba" sort="Chougrad, Hiba" uniqKey="Chougrad H" first="Hiba" last="Chougrad">Hiba Chougrad</name>
</noRegion>
<name sortKey="Alheyane, Omar" sort="Alheyane, Omar" uniqKey="Alheyane O" first="Omar" last="Alheyane">Omar Alheyane</name>
<name sortKey="Zouaki, Hamid" sort="Zouaki, Hamid" uniqKey="Zouaki H" first="Hamid" last="Zouaki">Hamid Zouaki</name>
</country>
</tree>
</affiliations>
</record>
Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Sante/explor/MaghrebDataLibMedV2/Data/Main/Exploration
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000442 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Exploration/biblio.hfd -nk 000442 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien |wiki= Wicri/Sante |area= MaghrebDataLibMedV2 |flux= Main |étape= Exploration |type= RBID |clé= pubmed:29477427 |texte= Deep Convolutional Neural Networks for breast cancer screening. }}
Pour générer des pages wiki
HfdIndexSelect -h $EXPLOR_AREA/Data/Main/Exploration/RBID.i -Sk "pubmed:29477427" \ | HfdSelect -Kh $EXPLOR_AREA/Data/Main/Exploration/biblio.hfd \ | NlmPubMed2Wicri -a MaghrebDataLibMedV2
This area was generated with Dilib version V0.6.38. |