articleIEEE Transactions on Medical ImagingFeb 22, 2019GREEN OA

Deep Learning for Segmentation Using an Open Large-Scale Dataset in 2D Echocardiography

SLSarah LeclercESErik SmistadJPJoao PedrosaAOAndreas OstvikFCFrederic Cervenansky

Université Claude Bernard Lyon 1 · Centre National de la Recherche Scientifique · +7 more institutions

PubMed
Indexed inarxivcrossrefpubmed

Abstract

Delineation of the cardiac structures from 2D echocardiographic images is a common clinical task to establish a diagnosis. Over the past decades, the automation of this task has been the subject of intense research. In this paper, we evaluate how far the state-of-the-art encoder-decoder deep convolutional neural network methods can go at assessing 2D echocardiographic images, i.e., segmenting cardiac structures and estimating clinical indices, on a dataset, especially, designed to answer this objective. We, therefore, introduce the cardiac acquisitions for multi-structure ultrasound segmentation dataset, the largest publicly-available and fully-annotated dataset for the purpose of echocardiographic assessment.…

Citation impact

689
total citations
FWCI
30.40
Percentile
100%
References
30
Citations per year

Authors

14
  • SL
    Sarah LeclercCorresponding

    Université Claude Bernard Lyon 1, Centre National de la Recherche Scientifique, Inserm, Centre de Recherche en Acquisition et Traitement de l'Image pour la Santé

  • ES
    Erik Smistad

    Norwegian University of Science and Technology

  • JP
    Joao Pedrosa

    KU Leuven

  • AO
    Andreas Ostvik
  • FC
    Frederic Cervenansky

    Université Claude Bernard Lyon 1, Centre National de la Recherche Scientifique, Inserm, Centre de Recherche en Acquisition et Traitement de l'Image pour la Santé

Topics & keywords

Keywords
  • Deep learning
  • Segmentation
  • Convolutional neural network
  • Mean absolute error
  • Pattern recognition (psychology)
  • Artificial neural network
  • Correlation coefficient
  • Ejection fraction
No related works found for this paper.

Funding