A Competitive Deep Neural Network Approach for the ImageCLEFmed Caption 2020 Task

TL;DR

This paper presents a deep neural network method for automatically labeling radiology images with medical concepts, achieving competitive results in the ImageCLEFmed Caption 2020 challenge without external data or pretrained models.

Contribution

The authors develop a DNN approach that effectively labels radiology images using only the provided data, demonstrating strong performance in a competitive challenge.

Findings

Achieved an F1 score of 0.375 on the test set

Ranked 12th among all submitted systems

Did not use external medical knowledge or pretrained models

Abstract

The aim of ImageCLEFmed Caption task is to develop a system that automatically labels radiology images with relevant medical concepts. We describe our Deep Neural Network (DNN) based approach for tackling this problem. On the challenge test set of 3,534 radiology images, our system achieves an F1 score of 0.375 and ranks high, 12th among all systems that were successfully submitted to the challenge, whereby we only rely on the provided data sources and do not use any external medical knowledge or ontologies, or pretrained models from other medical image repositories or application domains.