AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

RadEdit: stress-testing biomedical vision models via diffusion image editing

Fernando Pérez-García, Sam Bond-Taylor, Pedro Sanchez, Boris van Breugel, Daniel Coelho de Castro, Harshita Sharma, Valentina Salvatelli, Maria Teodora A Wetscherek, Hannah CM Richardson, Lungren Matthew, Aditya Nori, Javier Alvarez-Valle and 2 more

ECCV 2024 · Image generation › Text-to-image diffusion · approach: Text-driven synthesis and semantic guidance · 19 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Train segmentation on larger, abnormality-rich CheXmask instead of MIMIC-SegLung segmentation model · Healthy->pacemaker (RadEdit synthetic)Dice drop vs real data: 12.4 -> 8.2Improved
Train segmentation on CheXmask instead of MIMIC-SegLung segmentation model · Healthy->consolidation (RadEdit synthetic)Dice drop vs real data: 11.5 -> 7.4Improved
Train segmentation on CheXmask instead of MIMIC-SegLung segmentation model · Healthy->edema (RadEdit synthetic)Dice drop vs real data: 3.6 -> 1.6Improved
Use masks and DDPM inversion in RadEdit rather than LANCE or DDIM inversion (ablation)RadEdit editing · Lung border preservation after edema editDice / AHD of strong predictor vs original masksImproved
Use RadEdit keep/edit masks to remove COVID-19 features instead of unmasked editing (LANCE)COVID-19 acquisition shift editing · BIMCV+ editsPreservation of laterality markers and contrastImproved
Add a chest drain keep mask when editing out pneumothoraxRadEdit · CANDID-PTX editsChest drain preserved after editImproved
Use DiffEdit automatic masks instead of ground-truth masksPneumothorax mask prediction · Pneumothorax segmentationDice / AHDHurt
Strong predictor (trained on BIMCV+ and BIMCV-) tested on real vs synthetic edited setCOVID-19 classifier · BIMCV+/MIMIC-CXR vs RadEdit syntheticAccuracyNo clear effect
Weak COVID-19 classifier evaluated on RadEdit synthetic negativesWeak predictor (BIMCV+ vs MIMIC-CXR) · Synthetic COVID-negative imagesAccuracy: 95% dropHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy