English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

To effectively train medical students to become qualified radiologists, alarge number of X-ray images collected from patients with diverse medicalconditions are needed. However, due to data privacy concerns, such images aretypically difficult to obtain. To address this problem, we develop methods togenerate view-consistent, high-fidelity, and high-resolution X-ray images fromradiology reports to facilitate radiology training of medical students. Thistask is presented with several challenges. First, from a single report, imageswith different views (e.g., frontal, lateral) need to be generated. How toensure consistency of these images (i.e., make sure they are about the samepatient)? Second, X-ray images are required to have high resolution. Otherwise,many details of diseases would be lost. How to generate high-resolutionsimages? Third, radiology reports are long and have complicated structure. Howto effectively understand their semantics to generate high-fidelity images thataccurately reflect the contents of the reports? To address these threechallenges, we propose an XRayGAN composed of three modules: (1) a viewconsistency network that maximizes the consistency between generatedfrontal-view and lateral-view images; (2) a multi-scale conditional GAN thatprogressively generates a cascade of images with increasing resolution; (3) ahierarchical attentional encoder that learns the latent semantics of aradiology report by capturing its hierarchical linguistic structure and variouslevels of clinical importance of words and sentences. Experiments on tworadiology datasets demonstrate the effectiveness of our methods. To our bestknowledge, this work represents the first one generating consistent andhigh-resolution X-ray images from radiology reports. The code is available athttps://github.com/UCSD-AI4H/XRayGAN.

XRayGAN: Consistency-preserving Generation of X-ray Images from Radiology Reports