Article (Scientific journals)
Multi-task pre-training of deep neural networks for digital pathology
Mormont, Romain; Geurts, Pierre; Marée, Raphaël
2020In IEEE Journal of Biomedical and Health Informatics
Peer Reviewed verified by ORBi
 

Files


Full Text
mormont2020multitask.pdf
Author preprint (1.36 MB)
Download
Annexes
mormont2020multitask-supp.pdf
Publisher postprint (7.41 MB)
Supplementary materials
Download
isbi2021_abstract.pdf
Publisher postprint (124.88 kB)
Abstract ISBI 2021
Download

All documents in ORBi are protected by a user license.

Send to



Details



Keywords :
deep learning; multi-task learning; digital pathology; transfer learning
Abstract :
[en] In this work, we investigate multi-task learning as a way of pre-training models for classification tasks in digital pathology. It is motivated by the fact that many small and medium-size datasets have been released by the community over the years whereas there is no large scale dataset similar to ImageNet in the domain. We first assemble and transform many digital pathology datasets into a pool of 22 classification tasks and almost 900k images. Then, we propose a simple architecture and training scheme for creating a transferable model and a robust evaluation and selection protocol in order to evaluate our method. Depending on the target task, we show that our models used as feature extractors either improve significantly over ImageNet pre-trained models or provide comparable performance. Fine-tuning improves performance over feature extraction and is able to recover the lack of specificity of ImageNet features, as both pre-training sources yield comparable performance.
Disciplines :
Computer science
Author, co-author :
Mormont, Romain  ;  Université de Liège - ULiège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Algorith. des syst. en interaction avec le monde physique
Geurts, Pierre  ;  Université de Liège - ULiège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Algorith. des syst. en interaction avec le monde physique
Marée, Raphaël  ;  Université de Liège - ULiège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Méthodes stochastiques
Language :
English
Title :
Multi-task pre-training of deep neural networks for digital pathology
Alternative titles :
[fr] Pré-entraînement multi-tâche de réseaux de neurones pour la pathologie digitale
Publication date :
2020
Journal title :
IEEE Journal of Biomedical and Health Informatics
ISSN :
2168-2194
eISSN :
2168-2208
Publisher :
Institute of Electrical and Electronics Engineers, United States
Special issue title :
Special issue on Computational Pathology
Peer reviewed :
Peer Reviewed verified by ORBi
Available on ORBi :
since 04 May 2020

Statistics


Number of views
591 (50 by ULiège)
Number of downloads
1075 (25 by ULiège)

Scopus citations®
 
42
Scopus citations®
without self-citations
42
OpenCitations
 
8
OpenAlex citations
 
46

Bibliography


Similar publications



Contact ORBi