Research map / Label-efficient and robust learning / Recognition and long-tail
Incomplete Multi-view Diabetic Retinopathy Grading via Self-Supervised Inter- and Intra-View Restoration
Zhihao Wu, Yuxin Lin, Jie Wen, Wuzhen Shi, Linlin Shen
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 |
|---|---|---|---|
| add inter-view masking in SMCL (ablation) | Ours · Four-view DR grading (few views) | accuracy | Improved |
| add intra-view masking in SMCL (ablation) | Ours · Four-view DR grading | accuracy | Improved |
| use CSCA instead of conventional self-attention (ViT) module (ablation) | Ours · Four-view DR grading | accuracy | Improved |
| use 25% masking ratio (ablation) | Ours · Four-view DR grading | accuracy | Improved |
| use SMCL instead of DINO or SimMIM pretraining (ablation) | Swin-B · Four-view DR grading (full views) | accuracy: 81.7 / 82.9 -> 84.1 | Improved |
| use lighter backbone (ConvNeXt-Tiny / PVT-V2-B0) instead of Swin-B (ablation) | Ours · Four-view DR grading (full views) | accuracy: 84.1 -> 83.0 / 83.1 | Hurt |
| use Ours instead of WGLIMLs with one view | incomplete multi-view DR grading · Four-view DR grading (1 view) | accuracy: 78.4 -> 79.2 | Improved |
| use Ours instead of SMVDRLs with one view | incomplete multi-view DR grading · Four-view DR grading (1 view) | accuracy: 71.8 -> 79.2 | Improved |
| add lesion segmentation mask input to Ours | Ours · Four-view DR grading (4 views) | accuracy: 84.1 -> 84.7 | Improved |
| use Ours instead of WGLIMLs with all four views | multi-view DR grading · Four-view DR grading (4 views) | accuracy: 84.1 vs 84.1 | No clear effect |
Closest papers by meaning
- Like an Ophthalmologist: Dynamic Selection Driven Multi-View Learning for Diabetic Retinopathy Grading (AAAI 2025 · similarity 0.88)
- HACDR-Net: Heterogeneous-Aware Convolutional Network for Diabetic Retinopathy Multi-Lesion Segmentation (AAAI 2024 · similarity 0.86)
- FOCUS: Benchmarking Retinal Model Generalization from Foundation Vision Encoders to Multimodal LLMs (NeurIPS 2026 · similarity 0.86)
- Deep Hierarchies and Invariant Disease-Indicative Feature Learning for Computer Aided Diagnosis of Multiple Fundus Diseases (AAAI 2025 · similarity 0.85)
- SIGraph: Saliency Image-Graph Network for Retinal Disease Classification in Fundus Image (AAAI 2025 · similarity 0.83)
- Domain Adaptive Diabetic Retinopathy Grading with Model Absence and Flowing Data (CVPR 2025 · 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.
