Aleatoric uncertainty estimation with test-time augmentation for medical image segmentation with convolutional neural networks

Wang, Guotai; Li, Wenqi; Aertsen, Michaël; Deprest, Jan; Ourselin, Sébastien; Vercauteren, Tom

doi:10.1016/j.neucom.2019.01.103

articleNeurocomputingFeb 10, 2019HYBRID OA

Aleatoric uncertainty estimation with test-time augmentation for medical image segmentation with convolutional neural networks

GWGuotai Wang WLWenqi Li MAMichaël Aertsen JDJan Deprest SOSébastien Ourselin

University of Electronic Science and Technology of China · King's College London · +4 more institutions

PubMed

Indexed inarxivcrossrefpubmed

Abstract

Despite the state-of-the-art performance for medical image segmentation, deep convolutional neural networks (CNNs) have rarely provided uncertainty estimations regarding their segmentation outputs, e.g., model (epistemic) and image-based (aleatoric) uncertainties. In this work, we analyze these different types of uncertainties for CNN-based 2D and 3D medical image segmentation tasks at both pixel level and structure level. We additionally propose a test-time augmentation-based aleatoric uncertainty to analyze the effect of different transformations of the input image on the segmentation output. Test-time augmentation has been previously used to improve segmentation accuracy, yet not been formulated in a…

Citation impact

633

total citations

FWCI: 27.86
Percentile: 100%
References: 80

Citations per year

Authors

6

Topics & keywords

Topics

Keywords

Convolutional neural network
Artificial intelligence
Computer science
Pattern recognition (psychology)
Estimation
Image (mathematics)
Segmentation
Artificial neural network

No related works found for this paper.