Radgen: A Cross-Modal Fusion System for Automated Radiology Report Generation

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

Abstract

Automating radiology report generation can significantly reduce the workload of radiologists while improving the accuracy and consistency of clinical documentation. However, achieving optimal alignment between visual and textual representations in medical imaging remains a challenge. To address this, we demonstrate RadGen, a cross-modal fusion based system for automated medical report generation. RadGen uses MedCLIP as both a vision extractor and a retrieval mechanism to enhance the integration of imaging and textual data. By extracting features from retrieved reports and medical images through an attentionbased extraction module and integrating them with a fusion module, our system improves the coherence, accuracy, and clinical relevance of generated reports.

Original languageEnglish
Title of host publicationProceedings - 2025 IEEE 38th International Symposium on Computer-Based Medical Systems, CBMS 2025
EditorsAlejandro Rodriguez-Gonzalez, Rosa Sicilia, Lucia Prieto-Santamaria, George A. Papadopoulos, Valerio Guarrasi, Mirela Teixeira Cazzolato, Bridget Kane
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages111-112
Number of pages2
ISBN (Electronic)9798331526108
DOIs
StatePublished - 2025
Externally publishedYes
Event38th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2025 - Madrid, Spain
Duration: 18 Jun 202520 Jun 2025

Publication series

NameProceedings - IEEE Symposium on Computer-Based Medical Systems
ISSN (Print)1063-7125

Conference

Conference38th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2025
Country/TerritorySpain
CityMadrid
Period18/06/2520/06/25

Keywords

  • attention-based models
  • cross-modal
  • radiology report generation

Fingerprint

Dive into the research topics of 'Radgen: A Cross-Modal Fusion System for Automated Radiology Report Generation'. Together they form a unique fingerprint.

Cite this