.png){kind=icon}
.png){kind=icon}
Return
PiPViT: Patch-based Visual Interpretable Prototypes for retinal image analysis
.png){kind=icon}
OA
AI
DOI：10.1016/j.bspc.2026.109545.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• PiPViT is an interpretable prototypical model for retinal OCT slice classification. • It uses contrastive learning and multi-resolution inputs to localize biomarkers. • PiPViT estimates lesion extent using only image-level labels. • PiPViT learns prototypes that align with human-understandable biomarkers.
Keywords:
Interpretability
Prototype-based models
Vision transformers
Optical coherence tomography
Retina
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
4.9
Papers:
9.8K
Citations:
2.4W