Deep Quality Estimation: Creating Surrogate Models for Human Quality Ratings

Florian Kofler*, Ivan Ezhov, Lucas Fidon, Izabela Horvath, Ezequiel de la Rosa, John LaMaster, Hongwei Li, Tom Finck, Suprosanna Shit, Johannes Paetzold, Spyridon Bakas, Marie Piraud, Jan Kirschke, Tom Vercauteren, Claus Zimmer, Benedikt Wiestler, Bjoern Menze

*Corresponding author for this work

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

Abstract

Human ratings are abstract representations of segmentation quality. To approximate human quality ratings on scarce expert data, we train surrogate quality estimation models. We evaluate on a complex multi-class segmentation problem, specifically glioma segmentation, following the BraTS annotation protocol. The training data features quality ratings from 15 expert neuroradiologists on a scale ranging from 1 to 6 stars for various computer-generated and manual 3D annotations. Even though the networks operate on 2D images and with scarce training data, we can approximate segmentation quality within a margin of error comparable to human intra-rater reliability. Segmentation quality prediction has broad applications. While an understanding of segmentation quality is imperative for successful clinical translation of automatic segmentation quality algorithms, it can play an essential role in training new segmentation models. Due to the split-second inference times, it can be directly applied within a loss function or as a fully-automatic dataset curation mechanism in a federated learning setting.

Original languageEnglish
Title of host publicationBrainlesion
Subtitle of host publicationGlioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 8th International Workshop, BrainLes 2022, Held in Conjunction with MICCAI 2022, Revised Selected Papers
EditorsSpyridon Bakas, Ujjwal Baid, Bhakti Baheti, Alessandro Crimi, Sylwia Malec, Monika Pytlarz, Maximilian Zenk, Reuben Dorent
PublisherSpringer Science and Business Media Deutschland GmbH
Pages3-13
Number of pages11
ISBN (Print)9783031338410
DOIs
Publication statusPublished - 18 Jul 2023
EventProceedings of the 8th International MICCAI Brainlesion Workshop, BrainLes 2022 - Singapore, Singapore
Duration: 18 Sept 202222 Sept 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13769 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceProceedings of the 8th International MICCAI Brainlesion Workshop, BrainLes 2022
Country/TerritorySingapore
CitySingapore
Period18/09/202222/09/2022

Keywords

  • automatic quality control
  • BraTS
  • glioma
  • quality estimation
  • segmentation quality metrics

Fingerprint

Dive into the research topics of 'Deep Quality Estimation: Creating Surrogate Models for Human Quality Ratings'. Together they form a unique fingerprint.

Cite this