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

Radiology Report Generation (RRG) draws attention as an interaction betweenvision and language fields. Previous works inherited the ideology ofvision-to-language generation tasks,aiming to generate paragraphs with highconsistency as reports. However, one unique characteristic of RRG, theindependence between diseases, was neglected, leading to the injection ofdisease co-occurrence as a confounder that effects the results through backdoorpath. Unfortunately, this confounder confuses the process of report generationworse because of the biased RRG data distribution. In this paper, to rethinkthis issue thoroughly, we reason about its causes and effects from a novelperspective of statistics and causality, where the Joint Vision Coupling andthe Conditional Sentence Coherence Coupling are two aspects prone to implicitlydecrease the accuracy of reports. Then, a counterfactual augmentation strategythat contains the Counterfactual Sample Synthesis and the Counterfactual ReportReconstruction sub-methods is proposed to break these two aspects of spuriouseffects. Experimental results and further analyses on two widely used datasetsjustify our reasoning and proposed methods.

Rethinking Radiology Report Generation via Causal Reasoning and Counterfactual Augmentation