AI Paper Atlas

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

ChEX: Interactive Localization and Region Description in Chest X-rays

Philip Müller, Georgios Kaissis, Daniel Rueckert

ECCV 2024 · Image generation › Text-to-image diffusion · approach: T2I evaluation and controllable adapters · 20 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
add bounding box targets in multitask training (ablation)ChEX · MS-CXR sentence groundingmIoU: 3.68 -> 47.52Improved
add bounding box targets in multitask training (ablation)ChEX · MIMIC-CXR report generationMic-F1-14: 40.68 -> 52.32Improved
add post decoder in the sentence generator (ablation)ChEX · MIMIC-CXR report generationMic-F1-14: 46.07 -> 52.32Improved
add sentence tokens to the training mix (ablation)ChEX · MS-CXR sentence groundingmIoU: 52.97 -> 47.52Mixed
train with pathology, anatomy and sentence tokens together instead of only sentence tokens (ablation)ChEX · VinDr-CXR pathology detectionmAP: 0.06 -> 14.12Improved
add anatomy tokens to the training mix (ablation)ChEX · MIMIC-CXR report generationMic-F1-14: 46.64 -> 52.32Improved
add pathology classification labels (ablation)ChEX · VinDr-CXR pathology detectionmAP: 10.35 -> 14.12Improved
use precise regional hints in textual prompts (e.g. pneumonia in the left lung) (ablation)ChEX · pathology localization (gIoU)localization and prediction accuracyImproved
refine regional hint to finer region (upper lung) (ablation)ChEX · pathology localizationlocalization qualityNo clear effect
negate the prompt (ablation)ChEX · pathology localizationlocalization qualityHurt
use textual region prompts instead of bounding box queries (ablation)ChEX · Chest ImaGenome region explanationMic-F1-14: 53.34 -> 54.70Improved
skip decoder language model pretraining on MIMIC-CXR (ablation)ChEX sentence generator · MS-CXR region explanationMic-F1-14: 49.97 -> 48.56Hurt

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