Radiology report generation aims at generating descriptive text from radiology images automatically, which may present an opportunity to improve radiology reporting and interpretation. A typical setting consists of training encoder-decoder models on image-report pairs with a cross entropy loss, which struggles to generate informative sentences for clinical diagnoses since normal findings dominate the datasets. To tackle this challenge and encourage more clinically-accurate text outputs, we propose a novel weakly supervised contrastive loss for medical report generation. Experimental results demonstrate that our method benefits from contrasting target reports with incorrect but semantically-close ones. It outperforms previous work on both clinical correctness and text generation metrics for two public benchmarks.
@article{arxiv.2109.12242,
title = {Weakly Supervised Contrastive Learning for Chest X-Ray Report Generation},
author = {An Yan and Zexue He and Xing Lu and Jiang Du and Eric Chang and Amilcare Gentili and Julian McAuley and Chun-Nan Hsu},
journal= {arXiv preprint arXiv:2109.12242},
year = {2021}
}