Research map / Graph learning / Graph self-supervised learning
Vision-Language Neural Graph Featurization for Extracting Retinal Lesions
Taimur Hassan, Anabia Sohail, Muzammal Naseer, Naoufel Werghi
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 changed | Where | Measured | Effect |
|---|---|---|---|
| increase SLIC superpixels from 10 to 100 (ablation) | NGF · ODIR | F1: 0.79 -> 0.89 | Improved |
| increase SLIC superpixels from 10 to 100 (ablation) | NGF · BIOMISA | F1: 0.80 -> 0.93 | Improved |
| increase SLIC superpixels from 10 to 100 (ablation) | NGF · IDRiD | F1: 0.77 -> 0.92 | Improved |
| use PVT instead of ViT as node encoder (ablation) | NGF · ODIR | F1: 0.86 -> 0.89 | Improved |
| use PVT instead of ResNet-50 as node encoder (ablation) | NGF · BIOMISA | F1: 0.90 -> 0.93 | Improved |
| use PVT instead of ViT as node encoder (ablation) | NGF · IDRiD | F1: 0.90 -> 0.92 | Improved |
| use PVT instead of ResNet-50 as node encoder (ablation) | NGF · ODIR | time (ms): 462 -> 323 | Improved |
| fine-tune NGF models on 20 ODIR examples per class (few-shot) (ablation) | NGF zero-shot · ODIR | F1: 0.89 -> 0.94 | Improved |
| fine-tune NGF models on 20 ODIR examples per class (few-shot) (ablation) | NGF zero-shot · BIOMISA | F1: 0.93 -> 0.96 | Improved |
| fine-tune NGF models on 20 ODIR examples per class (few-shot) (ablation) | NGF zero-shot · IDRiD | IoU: 0.85 -> 0.90 | Improved |
| use unsupervised NGF instead of supervised DMSF | NGF zero-shot · ODIR | F1: 0.84 -> 0.89 | Improved |
| use unsupervised NGF instead of supervised DMSF | NGF zero-shot · BIOMISA | IoU: 0.83 -> 0.86 | Improved |
Closest papers by meaning
- SIGraph: Saliency Image-Graph Network for Retinal Disease Classification in Fundus Image (AAAI 2025 · similarity 0.87)
- FOCUS: Benchmarking Retinal Model Generalization from Foundation Vision Encoders to Multimodal LLMs (NeurIPS 2026 · similarity 0.87)
- HACDR-Net: Heterogeneous-Aware Convolutional Network for Diabetic Retinopathy Multi-Lesion Segmentation (AAAI 2024 · similarity 0.87)
- Unleashing the Potential of Vision-Language Pre-Training for 3D Zero-Shot Lesion Segmentation via Mask-Attribute Alignment (ICLR 2025 · similarity 0.84)
- EasyLens: A Training-Free Plug-and-Play Subtle-Lesion Representation Amplifier for Medical Vision-Language Models (NeurIPS 2026 · similarity 0.84)
- Like an Ophthalmologist: Dynamic Selection Driven Multi-View Learning for Diabetic Retinopathy Grading (AAAI 2025 · similarity 0.83)
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.
