Variational Topic Inference for Chest X-Ray Report Generation

Abstract

Automating report generation for medical imaging promises to reduce workload and assist diagnosis in clinical practice. Recent work has shown that deep learning models can successfully caption natural images. However, learning from medical data is challenging due to the diversity and uncertainty inherent in the reports written by different radiologists with discrepant expertise and experience. To tackle these challenges, we propose variational topic inference for automatic report generation. Specifically, we introduce a set of topics as latent variables to guide sentence generation by aligning image and language modalities in a latent space. The topics are inferred in a conditional variational inference framework, with each topic governing the generation of a sentence in the report. Further, we adopt a visual attention module that enables the model to attend to different locations in the image and generate more informative descriptions. We conduct extensive experiments on two benchmarks, namely Indiana U. Chest X-rays and MIMIC-CXR. The results demonstrate that our proposed variational topic inference method can generate novel reports rather than mere copies of reports used in training, while still achieving comparable performance to state-of-theart methods in terms of standard language generation criteria.

Reinforced Cross-modal Alignment for Radiology Report Generation

“…The task that is the most relevant to ours is image captioning, which aims to generation text captions that describe the content of the given images (Vinyals et al, 2015;Xu et al, 2015;Anderson et al, 2018;. Being one of its applications and extensions to the medical domain, radiology report generation aims to depicting radiology images with professional texts (Liu et al, 2019;Huang et al, 2019;Miura et al, 2020;Alfarghaly et al, 2021;Nooralahzadeh et al, 2021;Najdenkoska et al, 2021;Wang et al, 2021). In general, existing approaches for radiology report generation were mainly designed and proposed to better align images and texts or to exploit highly-patternized features of texts.…”