Reference : The effect of dataset confounding on predictions of deep neural networks for medical ...
Scientific congresses, symposiums and conference proceedings : Paper published in a book
Life sciences : Multidisciplinary, general & others
Engineering, computing & technology : Computer science
Engineering, computing & technology : Multidisciplinary, general & others
http://hdl.handle.net/10993/50409
The effect of dataset confounding on predictions of deep neural networks for medical imaging
English
Garcia Santa Cruz, Beatriz mailto [University of Luxembourg > Luxembourg Centre for Systems Biomedicine (LCSB) > >]
Husch, Andreas mailto [University of Luxembourg > Luxembourg Centre for Systems Biomedicine (LCSB) > Interventional Neuroscience >]
Hertel, Frank mailto [University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > >]
18-Apr-2022
Vol. 3 (2022): Proceedings of the Northern Lights Deep Learning Workshop 2022
8
Yes
No
International
4th Northern Lights Deep learning Conference 2022
from 10-1-2022 to 12-1-2022
[en] Confounding ; Deep learning ; medical imaging
[en] The use of Convolutional Neural Networks (CNN) in medical imaging has often outperformed previous solutions and even specialists, becoming a promising technology for Computer-aided-Diagnosis (CAD) systems. However, recent works suggested that CNN may have poor generalisation on new data, for instance, generated in different hospitals. Uncontrolled confounders have been proposed as a common reason. In this paper, we experimentally demonstrate the impact of confounding data in unknown scenarios. We assessed the effect of four confounding configurations: total, strong, light and balanced. We found the confounding effect is especially prominent in total confounder scenarios, while the effect on light and strong confounding scenarios may depend on the dataset robustness. Our findings indicate that the confounding effect is independent of the architecture employed.
These findings might explain why models can report good metrics during the development stage but fail to translate to real-world settings. We highlight the need for thorough consideration of these commonly unattended aspects, to develop safer CNN-based CAD systems.
Researchers ; Professionals ; Students
http://hdl.handle.net/10993/50409
10.7557/18.6302
https://septentrio.uit.no/index.php/nldl/article/view/6302

File(s) associated to this reference

Fulltext file(s):

FileCommentaryVersionSizeAccess
Open access
camara_ready_the_effect_of_dataset_confounding.pdfPublisher postprint458.5 kBView/Open

Bookmark and Share SFX Query

All documents in ORBilu are protected by a user license.