Unsupervised Domain Adaptation with Optimal Transport in multi-site segmentation of Multiple Sclerosis lesions from MRI data: Preprint

Antoine Ackaouy 1 Nicolas Courty 2, 3 Emmanuel Vallee 4 Olivier Commowick 1 Christian Barillot 1 Francesca Galassi 1
1 Empenn
IRISA-D5 - SIGNAUX ET IMAGES NUMÉRIQUES, ROBOTIQUE, Inria Rennes – Bretagne Atlantique , INSERM - Institut National de la Santé et de la Recherche Médicale
2 PANAMA - Parcimonie et Nouveaux Algorithmes pour le Signal et la Modélisation Audio
Inria Rennes – Bretagne Atlantique , IRISA-D5 - SIGNAUX ET IMAGES NUMÉRIQUES, ROBOTIQUE
3 OBELIX - Environment observation with complex imagery
UBS - Université de Bretagne Sud, IRISA-D5 - SIGNAUX ET IMAGES NUMÉRIQUES, ROBOTIQUE
Abstract : Automatic segmentation of Multiple Sclerosis (MS) lesions from Magnetic Resonance Imaging (MRI) images is essential for clinical assessment and treatment planning of MS. Recent years have seen an increasing use of Convolutional Neural Networks (CNNs) for this task. Although these methods provide accurate segmentation, their applicability in clinical settings remains limited due to a reproducibility issue across different image domains. MS images can have highly variable characteristics across patients, MRI scanners and imaging protocols; retraining a supervised model with data from each new domain is not a feasible solution because it requires manual annotation from expert radiologists. In this work, we explore an unsupervised solution to the problem of domain shift. We present a framework, Seg-JDOT, which adapts a deep model so that samples from a source domain and samples from a target domain sharing similar representations will be similarly segmented. We evaluated the framework on a multi-site dataset, MICCAI 2016, and showed that the adaptation towards a target site can bring remarkable improvements in a model performance over standard training.
Complete list of metadatas

Cited literature [37 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02317028
Contributor : Francesca Galassi <>
Submitted on : Tuesday, October 15, 2019 - 5:40:58 PM
Last modification on : Sunday, October 20, 2019 - 6:27:26 AM

File

preprint_ack.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02317028, version 1

Citation

Antoine Ackaouy, Nicolas Courty, Emmanuel Vallee, Olivier Commowick, Christian Barillot, et al.. Unsupervised Domain Adaptation with Optimal Transport in multi-site segmentation of Multiple Sclerosis lesions from MRI data: Preprint. 2019. ⟨hal-02317028⟩

Share

Metrics

Record views

221

Files downloads

83