Self-supervised Multi-scale Consistency for Weakly Supervised Segmentation Learning

Gabriele Valvano*, Andrea Leo, Sotirios A. Tsaftaris

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Collecting large-scale medical datasets with fine-grained annotations is time-consuming and requires experts. For this reason, weakly supervised learning aims at optimising machine learning models using weaker forms of annotations, such as scribbles, which are easier and faster to collect. Unfortunately, training with weak labels is challenging and needs regularisation. Herein, we introduce a novel self-supervised multi-scale consistency loss, which, coupled with an attention mechanism, encourages the segmentor to learn multi-scale relationships between objects and improves performance. We show state-of-the-art performance on several medical and non-medical datasets. The code used for the experiments is available at https://vios-s.github.io/multiscale-pyag.

Original languageEnglish
Title of host publicationDomain Adaptation and Representation Transfer, and Affordable Healthcare and AI for Resource Diverse Global Health
Subtitle of host publication3rd MICCAI Workshop, DART 2021, and 1st MICCAI Workshop, FAIR 2021, Held in Conjunction with MICCAI 2021, Proceedings
EditorsShadi Albarqouni, M. Jorge Cardoso, Qi Dou, Konstantinos Kamnitsas, Bishesh Khanal, Islem Rekik, Nicola Rieke, Debdoot Sheet, Sotirios Tsaftaris, Daguang Xu, Ziyue Xu
PublisherSpringer Science and Business Media Deutschland GmbH
Pages14-24
Number of pages11
ISBN (Print)9783030877217
DOIs
Publication statusPublished - 21 Sep 2021
Event3rd MICCAI Workshop on Domain Adaptation and Representation Transfer, DART 2021, and the 1st MICCAI Workshop on Affordable Healthcare and AI for Resource Diverse Global Health, FAIR 2021, held in conjunction with 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021 - Virtual, Online
Duration: 27 Sep 20211 Oct 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12968 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference3rd MICCAI Workshop on Domain Adaptation and Representation Transfer, DART 2021, and the 1st MICCAI Workshop on Affordable Healthcare and AI for Resource Diverse Global Health, FAIR 2021, held in conjunction with 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021
CityVirtual, Online
Period27/09/211/10/21

Keywords

  • Segmentation
  • Self-supervised learning
  • Shape prior

Fingerprint

Dive into the research topics of 'Self-supervised Multi-scale Consistency for Weakly Supervised Segmentation Learning'. Together they form a unique fingerprint.

Cite this