Reducing Textural Bias Improves Robustness of Deep Segmentation Models

Seoin Chai, Daniel Rueckert, Ahmed E. Fetit

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

Despite advances in deep learning, robustness under domain shift remains a major bottleneck in medical imaging settings. Findings on natural images suggest that deep neural models can show a strong textural bias when carrying out image classification tasks. In this thorough empirical study, we draw inspiration from findings on natural images and investigate ways in which addressing the textural bias phenomenon could bring up the robustness of deep segmentation models when applied to three-dimensional (3D) medical data. To achieve this, publicly available MRI scans from the Developing Human Connectome Project are used to study ways in which simulating textural noise can help train robust models in a complex semantic segmentation task. We contribute an extensive empirical investigation consisting of 176 experiments and illustrate how applying specific types of simulated textural noise prior to training can lead to texture invariant models, resulting in improved robustness when segmenting scans corrupted by previously unseen noise types and levels.

Original languageEnglish
Title of host publicationMedical Image Understanding and Analysis - 25th Annual Conference, MIUA 2021, Proceedings
EditorsBartłomiej W. Papież, Mohammad Yaqub, Jianbo Jiao, Ana I. Namburete, J. Alison Noble
PublisherSpringer Science and Business Media Deutschland GmbH
Pages294-304
Number of pages11
ISBN (Print)9783030804312
DOIs
StatePublished - 2021
Externally publishedYes
Event25th Annual Conference on Medical Image Understanding and Analysis, MIUA 2021 - Virtual, Online
Duration: 12 Jul 202114 Jul 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12722 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference25th Annual Conference on Medical Image Understanding and Analysis, MIUA 2021
CityVirtual, Online
Period12/07/2114/07/21

Keywords

  • Domain shift
  • Robustness
  • Segmentation
  • Textural bias

Fingerprint

Dive into the research topics of 'Reducing Textural Bias Improves Robustness of Deep Segmentation Models'. Together they form a unique fingerprint.

Cite this