AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

From Panel to Pixel: Zoom-In Vision-Language Pretraining from Biomedical Scientific Literature

Kun Yuan, Min Sun, Zhen Chen, Alejandro Lozano, Xiangteng He, Shi Li, Nassir Navab, Xiaoxiao Sun, Nicolas Padoy, Serena Yeung-Levy

CVPR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Recaptioning and text-supervised pretraining · 0 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 multi-granularity (figure, panel, region) supervision from Panel2PatchBMC-CLIP ViT-L/14 · Single-panel short-context retrieval, I2T R@10R@10 (%): 73.53 -> 79.90Improved
add region-level supervision to single-panel data (full Panel2Patch vs single-panel only) (ablation)Panel2Patch · Bounding-box to text retrieval, I2T R@5R@5 (%): 17.65 -> 20.24Improved
add single-panel data to region-only training (ablation)Panel2Patch · Single-panel short-context retrieval, I2T R@5R@5 (%): 61.60 -> 68.14Improved
add alternating coarse-to-fine training instead of a single level only (ablation)Panel2Patch · Multi-panel retrieval, 100K pairs, I2T R@5R@5 (%): 21.54 -> 22.07Improved
alternating training vs single-panel-only on single-panel test (ablation)Panel2Patch · Single-panel retrieval, 100K pairs, I2T R@10R@10 (%): 15.36 -> 14.79No clear effect
train on multi-panel only then test on single-panel (no alternating) (ablation)Panel2Patch · Cross-level retrieval, I2T R@10R@10 (%): Multi->Single 8.14 vs Multi->Multi 29.35Hurt
train on single-panel only then test on multi-panel (no alternating) (ablation)Panel2Patch · Cross-level retrieval, I2T R@10R@10 (%): Single->Multi 12.37 vs Ours->Multi 29.59Hurt
train on 400K hierarchical pairs instead of much larger corporaBMC-CLIP ViT-L/14 · Zero-shot classification, 6 specialties (avg)accuracy (avg): 47.85 -> 50.25Improved
Panel2Patch vs BMC-LongCLIP on Ophthalmology zero-shotBMC-CLIP ViT-L/14 · Zero-shot classification (Ophthalmology)accuracy: 59.80 -> 35.78Hurt
Panel2Patch on bounding-box text retrieval vs BMC-CLIP at higher recallBMC-CLIP ViT-L/14 · Bounding-box to text retrieval, I2T R@5R@5 (%): 20.98 -> 20.24Mixed

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