Research map / Label-efficient and robust learning / Recognition and long-tail
Post-training Feature Pruning for Fundus Images Classification
Van-Nguyen Pham, Duc-Tai Le, Junghyun Bum, Hyunseung Choo
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 |
|---|---|---|---|
| apply GFP post-training feature pruning to CoAtNet | CoAtNet-2 · DDR | AUROC: 95.17 -> 96.73 | Improved |
| apply GFP to CoAtNet | CoAtNet-2 · PAPILA | AUPRC: 80.70 -> 84.43 | Improved |
| apply GFP to EfficientNetV2 | EfficientNetV2-m · RETINA | AUROC: 77.85 -> 79.10 | Improved |
| apply GFP to EfficientNetV2 | EfficientNetV2-m · PAPILA | AUROC: 81.33 -> 81.21 | No clear effect |
| apply GFP to ViT | ViT · PAPILA | AUPRC: 79.27 -> 77.43 | Hurt |
| apply GFP to ViT | ViT · Messidor-2 | AUROC: 85.29 -> 85.36 | No clear effect |
| use magnitude pruning (MP) instead of GFP on EfficientNetV2 | EfficientNetV2-m · RETINA | AUROC: 77.85 -> 73.57 | Hurt |
| use L1 regularization feature selection on CoAtNet | CoAtNet-2 · DDR | AUROC: 95.17 -> 91.33 | Hurt |
| token pruning (TRAM, LTMP) instead of feature pruning on ViT | ViT · RETINA | AUROC: 90.75 -> 83.17 | Hurt |
| apply GFP to CoAtNet (feature compactness) | CoAtNet-2 · PAPILA | intra-class variance: 196.04 -> 8.41 | Improved |
| apply GFP to EfficientNetV2 for cross-dataset transfer | EfficientNetV2-m · Messidor-2 to DDR | AUROC: 80.83 -> 82.99 | Improved |
| apply GFP to EfficientNetV2 for cross-dataset transfer | EfficientNetV2-m · DDR to Messidor-2 | AUROC: 77.38 -> 77.17 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TRAM — TRAM: Finetuning-Free Test-Time Adaptation for Generalized Face Anti-Spoofing with Only a Few Bonafide Samples (ECCV 2026) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Deep Hierarchies and Invariant Disease-Indicative Feature Learning for Computer Aided Diagnosis of Multiple Fundus Diseases (AAAI 2025 · similarity 0.86)
- Attention Based Glaucoma Detection: A Large-Scale Database and CNN Model (CVPR 2019 · similarity 0.85)
- FOCUS: Benchmarking Retinal Model Generalization from Foundation Vision Encoders to Multimodal LLMs (NeurIPS 2026 · similarity 0.85)
- Like an Ophthalmologist: Dynamic Selection Driven Multi-View Learning for Diabetic Retinopathy Grading (AAAI 2025 · similarity 0.82)
- SIGraph: Saliency Image-Graph Network for Retinal Disease Classification in Fundus Image (AAAI 2025 · similarity 0.82)
- Optimal Eye Surgeon: Finding image priors through sparse generators at initialization (ICML 2024 · similarity 0.82)
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.
