GIT-CXR: End-to-End Transformer for Chest X-Ray Report Generation

Medical imaging is crucial for diagnosing, monitoring, and treating medical conditions. The medical reports of radiology images are the primary medium through which medical professionals can attest to their findings, but their writing is time-consuming and requires specialized clinical expertise. Th...

Full description

Saved in:
Bibliographic Details
Main Authors: Iustin Sîrbu, Iulia-Renata Sîrbu, Jasmina Bogojeska, Traian Rebedea
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:Information
Subjects:
Online Access:https://www.mdpi.com/2078-2489/16/7/524
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850077835734024192
author Iustin Sîrbu
Iulia-Renata Sîrbu
Jasmina Bogojeska
Traian Rebedea
author_facet Iustin Sîrbu
Iulia-Renata Sîrbu
Jasmina Bogojeska
Traian Rebedea
author_sort Iustin Sîrbu
collection DOAJ
description Medical imaging is crucial for diagnosing, monitoring, and treating medical conditions. The medical reports of radiology images are the primary medium through which medical professionals can attest to their findings, but their writing is time-consuming and requires specialized clinical expertise. Therefore, the automated generation of radiography reports has the potential to improve and standardize patient care and significantly reduce the workload of clinicians. Through our work, we have designed and evaluated an end-to-end transformer-based method to generate accurate and factually complete radiology reports for X-ray images. Additionally, we are the first to introduce curriculum learning for end-to-end transformers in medical imaging and demonstrate its impact in obtaining improved performance. The experiments were conducted using the MIMIC-CXR-JPG database, the largest available chest X-ray dataset. The results obtained are comparable with the current state of the art on the natural language generation (NLG) metrics BLEU and ROUGE-L, while setting new state-of-the-art results on F1 examples-averaged F1-macro and F1-micro metrics for clinical accuracy and on the METEOR metric widely used for NLG.
format Article
id doaj-art-6dcede9afabd4f9fa079a59653ce9678
institution DOAJ
issn 2078-2489
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series Information
spelling doaj-art-6dcede9afabd4f9fa079a59653ce96782025-08-20T02:45:42ZengMDPI AGInformation2078-24892025-06-0116752410.3390/info16070524GIT-CXR: End-to-End Transformer for Chest X-Ray Report GenerationIustin Sîrbu0Iulia-Renata Sîrbu1Jasmina Bogojeska2Traian Rebedea3Faculty of Automatic Control and Computer Science, National University of Science and Technology Politehnica of Bucharest, 060042 Bucharest, RomaniaFaculty of Automatic Control and Computer Science, National University of Science and Technology Politehnica of Bucharest, 060042 Bucharest, RomaniaSchool of Engineering, Zurich University of Applied Sciences, 8401 Winterthur, SwitzerlandFaculty of Automatic Control and Computer Science, National University of Science and Technology Politehnica of Bucharest, 060042 Bucharest, RomaniaMedical imaging is crucial for diagnosing, monitoring, and treating medical conditions. The medical reports of radiology images are the primary medium through which medical professionals can attest to their findings, but their writing is time-consuming and requires specialized clinical expertise. Therefore, the automated generation of radiography reports has the potential to improve and standardize patient care and significantly reduce the workload of clinicians. Through our work, we have designed and evaluated an end-to-end transformer-based method to generate accurate and factually complete radiology reports for X-ray images. Additionally, we are the first to introduce curriculum learning for end-to-end transformers in medical imaging and demonstrate its impact in obtaining improved performance. The experiments were conducted using the MIMIC-CXR-JPG database, the largest available chest X-ray dataset. The results obtained are comparable with the current state of the art on the natural language generation (NLG) metrics BLEU and ROUGE-L, while setting new state-of-the-art results on F1 examples-averaged F1-macro and F1-micro metrics for clinical accuracy and on the METEOR metric widely used for NLG.https://www.mdpi.com/2078-2489/16/7/524radiology report generationcurriculum learningimage captioningchest X-raytransformermachine learning
spellingShingle Iustin Sîrbu
Iulia-Renata Sîrbu
Jasmina Bogojeska
Traian Rebedea
GIT-CXR: End-to-End Transformer for Chest X-Ray Report Generation
Information
radiology report generation
curriculum learning
image captioning
chest X-ray
transformer
machine learning
title GIT-CXR: End-to-End Transformer for Chest X-Ray Report Generation
title_full GIT-CXR: End-to-End Transformer for Chest X-Ray Report Generation
title_fullStr GIT-CXR: End-to-End Transformer for Chest X-Ray Report Generation
title_full_unstemmed GIT-CXR: End-to-End Transformer for Chest X-Ray Report Generation
title_short GIT-CXR: End-to-End Transformer for Chest X-Ray Report Generation
title_sort git cxr end to end transformer for chest x ray report generation
topic radiology report generation
curriculum learning
image captioning
chest X-ray
transformer
machine learning
url https://www.mdpi.com/2078-2489/16/7/524
work_keys_str_mv AT iustinsirbu gitcxrendtoendtransformerforchestxrayreportgeneration
AT iuliarenatasirbu gitcxrendtoendtransformerforchestxrayreportgeneration
AT jasminabogojeska gitcxrendtoendtransformerforchestxrayreportgeneration
AT traianrebedea gitcxrendtoendtransformerforchestxrayreportgeneration