The automatic generation of radiology reports has the potential to assist radiologists in the time-consuming task of report writing. Existing methods generate the full report from image-level features, failing to explicitly focus on anatomical regions in the image. We propose a simple yet effective region-guided report generation model that detects anatomical regions and then describes individual, salient regions to form the final report. While previous methods generate reports without the possibility of human intervention and with limited explainability, our method opens up novel clinical use cases through additional interactive capabilities and introduces a high degree of transparency and explainability. Comprehensive experiments demonstrate our method's effectiveness in report generation, outperforming previous state-of-the-art models, and highlight its interactive capabilities. The code and checkpoints are available at https://github.com/ttanida/rgrg .
@article{arxiv.2304.08295,
title = {Interactive and Explainable Region-guided Radiology Report Generation},
author = {Tim Tanida and Philip Müller and Georgios Kaissis and Daniel Rueckert},
journal= {arXiv preprint arXiv:2304.08295},
year = {2023}
}