Imperial College London
Portrait Picture

Dr Ben Glocker

Faculty of Engineering, Department of Computing

Professor in Machine Learning for Imaging
 
 
 
//

Contact

 

+44 (0)20 7594 8334b.glocker Website CV

 
 
//

Location

 

377Huxley BuildingSouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@inproceedings{Kamnitsas:2021:10.1007/978-3-030-87722-4_8,
author = {Kamnitsas, K and Winzeck, S and Kornaropoulos, EN and Whitehouse, D and Englman, C and Phyu, P and Pao, N and Menon, DK and Rueckert, D and Das, T and Newcombe, VFJ and Glocker, B},
doi = {10.1007/978-3-030-87722-4_8},
pages = {79--89},
publisher = {Springer},
title = {Transductive image segmentation: Self-training and effect of uncertainty estimation},
url = {http://dx.doi.org/10.1007/978-3-030-87722-4_8},
year = {2021}
}
Download

RIS format (EndNote, RefMan)

TY  - CPAPER
AB  - Semi-supervised learning (SSL) uses unlabeled data during training to learnbetter models. Previous studies on SSL for medical image segmentation focusedmostly on improving model generalization to unseen data. In some applications,however, our primary interest is not generalization but to obtain optimalpredictions on a specific unlabeled database that is fully available duringmodel development. Examples include population studies for extracting imagingphenotypes. This work investigates an often overlooked aspect of SSL,transduction. It focuses on the quality of predictions made on the unlabeleddata of interest when they are included for optimization during training,rather than improving generalization. We focus on the self-training frameworkand explore its potential for transduction. We analyze it through the lens ofInformation Gain and reveal that learning benefits from the use of calibratedor under-confident models. Our extensive experiments on a large MRI databasefor multi-class segmentation of traumatic brain lesions shows promising resultswhen comparing transductive with inductive predictions. We believe this studywill inspire further research on transductive learning, a well-suited paradigmfor medical image analysis.
AU  - Kamnitsas,K
AU  - Winzeck,S
AU  - Kornaropoulos,EN
AU  - Whitehouse,D
AU  - Englman,C
AU  - Phyu,P
AU  - Pao,N
AU  - Menon,DK
AU  - Rueckert,D
AU  - Das,T
AU  - Newcombe,VFJ
AU  - Glocker,B
DO  - 10.1007/978-3-030-87722-4_8
EP  - 89
PB  - Springer
PY  - 2021///
SP  - 79
TI  - Transductive image segmentation: Self-training and effect of uncertainty estimation
UR  - http://dx.doi.org/10.1007/978-3-030-87722-4_8
UR  - http://arxiv.org/abs/2107.08964v1
UR  - https://link.springer.com/chapter/10.1007/978-3-030-87722-4_8
UR  - http://hdl.handle.net/10044/1/90453
ER  -
Download