Structured Knowledge Graphs for Classifying Unseen Patterns in Radiographs

Chinmay Prabhakar, Anjany Sekuboyina, Hongwei Bran Li, Johannes C. Paetzold, Suprosanna Shit, Tamaz Amiranashvili, Jens Kleesiek, Bjoern Menze

Research output: Contribution to journalConference articlepeer-review

Abstract

The presence of annotated datasets is crucial to the performance of modern machine learning algorithms. However, obtaining richly annotated datasets is not always possible, especially for novel or rare diseases. This becomes especially challenging in the realm of multi-label classification of chest radiographs, due to the presence of numerous unknown disease types and the limited information inherent to x-ray images. Ideally, we would like to develop models that can reliably label such unseen patterns (classes). In this work, we present a knowledge graph-based approach to predict such novel, unseen classes. Our method directly injects the semantic relationships between seen and unseen disease classes. Specifically, we propose a principled approach to parsing and processing a knowledge graph conditioned on the given task. We show that our method matches the labeling performance of the state-of-The-Art while outperforming it on unseen classes by a substantial 2% gain on chest X-ray classification. Crucially, we demonstrate that embedding diseasespecific knowledge as a graph provides inherent explainability. (The code is available at https://github.com/chinmay5/ml-cxr-gzsl-kg).

Original languageEnglish
Pages (from-to)45-60
Number of pages16
JournalProceedings of Machine Learning Research
Volume194
StatePublished - 2022
Externally publishedYes
Event1st International Workshop on Geometric Deep Learning in Medical Image Analysis, GeoMedIA 2022 - Amsterdam, Netherlands
Duration: 18 Nov 2022 → …

Keywords

  • generalized zero-shot learning
  • graph neural networks
  • knowledge graphs

Fingerprint

Dive into the research topics of 'Structured Knowledge Graphs for Classifying Unseen Patterns in Radiographs'. Together they form a unique fingerprint.

Cite this