Doubly Weak Supervision of Deep Learning Models for Head CT

Khaled Saab, Jared Dunnmon, Roger Goldman, Alex Ratner, Hersh Sagreiya, Christopher Ré, Daniel Rubin

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

5 Scopus citations

Abstract

Recent deep learning models for intracranial hemorrhage (ICH) detection on computed tomography of the head have relied upon large datasets hand-labeled at either the full-scan level or at the individual slice-level. Though these models have demonstrated favorable empirical performance, the hand-labeled datasets upon which they rely are time-consuming and expensive to create. Further, given limited time, modelers must currently make an explicit choice between scan-level supervision, which leverages large numbers of patients, and slice-level supervision, which yields clinically insightful output in the axial and in-plane dimensions. In this work, we propose doubly weak supervision, where we (1) weakly label at the scan-level to scalably incorporate data from large populations and (2) model the problem using an attention-based multiple-instance learning approach that can provide useful signal at both axial and in-plane granularities, even with scan-level supervision. Models trained using this doubly weak supervision approach yield an average ROC-AUC score of 0.91, which is competitive with those of models trained using large, hand-labeled datasets, while requiring less than 10 h of clinician labeling time. Further, our models place large attention weights on the same slices used by the clinician to arrive at the ICH classification, and occlusion maps indicate heavy influence from clinically salient in-plane regions.

Original languageEnglish (US)
Title of host publicationMedical Image Computing and Computer Assisted Intervention – MICCAI 2019 - 22nd International Conference, Proceedings
EditorsDinggang Shen, Pew-Thian Yap, Tianming Liu, Terry M. Peters, Ali Khan, Lawrence H. Staib, Caroline Essert, Sean Zhou
PublisherSpringer
Pages811-819
Number of pages9
ISBN (Print)9783030322472
DOIs
StatePublished - 2019
Externally publishedYes
Event22nd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2019 - Shenzhen, China
Duration: Oct 13 2019Oct 17 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11766 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference22nd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2019
Country/TerritoryChina
CityShenzhen
Period10/13/1910/17/19

Keywords

  • Head CT
  • Multiple instance learning
  • Weak supervision

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Doubly Weak Supervision of Deep Learning Models for Head CT'. Together they form a unique fingerprint.

Cite this