Skip to Main content Skip to Navigation
Conference papers

Une stratégie de données efficace pour la détection des anévrismes cérébraux avec l'apprentissage profond

youssef Assis 1 Liang Liao 2 Fabien Pierre 1 René Anxionnat 2, 3 Erwan Kerrien 1 
1 TANGRAM - Recalage visuel avec des modèles physiquement réalistes
Inria Nancy - Grand Est, UL - Université de Lorraine, LORIA - ALGO - Department of Algorithms, Computation, Image and Geometry, CNRS - Centre National de la Recherche Scientifique
Abstract : The detection of intracranial aneurysms from Magnetic Resonance Angiography images is a problem of rapidly growing clinical importance, but also extremely challenging to automate. However, in the last 3 years, the raise of deep convolutional neural networks has instigated a streak ofmethods that have convincingly removed the technological deadlock and show promising performance. The major issue to address is the very severe class imbalance. Previous authors have focused their efforts on the network architecture and loss function. This paper tackles the data. A rough but fast annotation is considered : each aneurysm is approximated by a sphere defined by two points. Second, a small patch approach is taken so as to increase the number of samples. Third, samples are generated by a combination of data selection (negative patches are centered half on blood vessels and half on parenchyma) and data synthesis (patches containing an aneurysm are duplicated and deformed by a 3D spline transform). This strategy is applied to train a 3D U-net model, with a binary cross entropy loss, on a data set of 111 patients. A 5-fold cross-validation evaluation provides state of the art results (sensitivity 0.82, false positive count 0.61, as per ADAM challenge criteria). The study also reports a comparison with the focal loss, and Cohen’s Kappa coefficient is shown to be a better metric than Dice for this highly unbalanced detection problem.
Complete list of metadata
Contributor : Youssef ASSIS Connect in order to contact the contributor
Submitted on : Wednesday, January 5, 2022 - 5:23:57 PM
Last modification on : Thursday, May 5, 2022 - 10:19:03 AM


Files produced by the author(s)


  • HAL Id : hal-03339672, version 2


youssef Assis, Liang Liao, Fabien Pierre, René Anxionnat, Erwan Kerrien. Une stratégie de données efficace pour la détection des anévrismes cérébraux avec l'apprentissage profond. ORASIS 2021 - 18èmes journées francophones des jeunes chercheurs en vision par ordinateur, Centre National de la Recherche Scientifique [CNRS], Sep 2021, Saint Ferréol, France. ⟨hal-03339672v2⟩



Record views


Files downloads