SAMOVAR UMR 5157

  • Accueil
  • Accueil
  • Accueil
  • Accueil

CNRS

Rechercher




Accueil > Équipes > ARMEDIA > Séminaires ARMEDIA > Séminaires 2018

Séminaire ARMEDIA, présenté par Sebastian Tarando le 5/04/18 à 14h en H218

Séminaire ARMEDIA, présenté par Sebastian Tarando le 5/04/18 à 14h en H218

L’équipe ARMEDIA organise un séminaire sur un thème en imagerie médicale, le jeudi 5 avril à 14h, salle H218 (site d’Evry) et salle visio de NanoInnov.
L’exposé suivant sera donné par Sebastian Tarando, doctorant ARMEDIA en 3è année.

Titre : " Deep learning framework for infiltrative lung disease classification "

Résumé :

Infiltrative lung diseases enclose a large group of irreversible lung disorders which require regular follow-up with CT Imaging. A quantitative assessment is mandatory to establish the (regional) disease progression and/or the therapeutic impact. This implies the development of automated computer-aided diagnosis (CAD) tools for pathological lung tissue segmentation, problem addressed as pixel-based texture classification. For the large majority of computer-aided diagnosis (CAD) systems, such classification relies on a two-dimensional analysis of axial CT images.

The inherent issue of ILD classification task is the great overlap between the texture pattern appearance which often makes the task an ontological problem leading to errors in the classification, i.e., mixing different types of lung patterns or overestimating the pathological areas. Recently, the use of deep leaning techniques, especially Convolutional Neural Networks (CNNs), has shown great improvements with respect to traditional methods (based on handcrafted features), emerging as the new state of the art for visual tasks and the reference for further comparisons with new approaches. Deep learning methods have the advantage (over traditional CAD approaches) that the features on which relies the classification are learned directly from the image dataset. This leads apparently to a better image pattern discrimination than when using handcrafted features. Note that CNNs particularly have proven to succeed in natural image classification tasks. However, their success on texture recognition problems has not been yet achieved (due to the higher intrinsic dimensionality of the latter datasets).

In this study we investigate the performance a deep learning approach for ILD classification. The proposed approach exploits a cascade of convolutional neural networks (specially designed for texture recognition) and a specific preprocessing of input data based on locally connected filtering. The classification targeting the whole lung parenchyma achieves an average of 84% accuracy (75.8% for normal, 90% for emphysema and fibrosis, 81.5% for ground glass).

The talk is concluded with a critical analysis of the presented approach and an outline of future research directions, followed by an open discussion with the audience.