AI Paper Atlas

Research map / Vision-language and multimodal / Visual question answering

Uncertainty-Aware Medical Diagnostic Phrase Identification and Grounding

Ke Zou, Yang Bai, Bo Liu, Yidi Chen, Zhihao Chen, Yang Zhou, Xuedong Yuan, Meng Wang, Xiaojing Shen, Xiaochun Cao, Yih Chung Tham, Huazhu Fu

TPAMI 2025 · Vision-language and multimodal › Visual question answering · approach: Multimodal LLM question answering · 39 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 uncertainty-aware prediction model (multi-hypothesis head) (ablation)MedGround · MRG-MS-CXRmIOU: 46.04 -> 47.65Improved
add uncertainty-aware prediction model (ablation)MedGround · MRG-ChestX-ray8AP50: 31.82 -> 38.38Improved
add SAM decoder fine-tuning with frozen encoder to LLaVA box regression (ablation)LLaVA+LoRA · MRG-MS-CXRAP30: 29.34 -> 71.26Improved
fine-tune SAM encoder with LoRA instead of keeping it frozen (ablation)uMedGround · MRG-MS-CXRmIOU: 47.65 -> 45.31Hurt
fully fine-tune SAM encoder and decoder instead of decoder only (ablation)uMedGround · MRG-MS-CXRmIOU: 47.65 -> 41.53Hurt
use LoRA-fine-tuned LLaMA2 phrases instead of zero-shot GPT-4 phrasesTransVG · MRG-MS-CXRmIOU: 26.28 -> 42.60Improved
replace two-stage phrase extraction with end-to-end uMedGroundTransVG with LLaMA2-FT phrases · MRG-MS-CXRmIOU: 42.60 -> 47.65Improved
add uncertainty-aware prediction under Gaussian noise 0.1 (ablation)MedGround · MRG-ChestX-ray8 (noisy)mIOU: 32.52 -> 35.16Improved
add Gaussian noise to input imagesuMedGround · MRG-MS-CXRmIOU: 47.65 -> 47.01Hurt
cross-dataset transfer: add uncertainty-aware prediction (ablation)MedGround · MRG-ChestX-ray8 (trained on MRG-MS-CXR)AP50: 21.72 -> 23.74Mixed
use VQA pairs instead of class descriptions as text inputuMedGround · MRG-MIMICmIOU: 46.61 -> 47.39Improved
use original report instead of VQA pairs as text inputuMedGround · MRG-MIMICmIOU: 47.39 -> 47.65Improved

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