Research map / Image generation / GANs
K-SALSA: K-Anonymous Synthetic Averaging of Retinal Images via Local Style Alignment
Minkyu Jeon, Hyeonjin Park, Hyunwoo J. Kim, Michael Morley, Hyunghoon Cho
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 local style alignment instead of Euclidean averaging | k-Centroid · APTOS k=5 | FID: 22.71 -> 20.09 | Mixed |
| add local style alignment instead of Euclidean averaging | k-Centroid · EyePACS k=10 | FID: 28.49 -> 21.28 | Mixed |
| add local style alignment instead of Euclidean averaging | k-Centroid · APTOS k=5 | accuracy: 0.688 -> 0.712 | Improved |
| add local style alignment instead of Euclidean averaging | k-Centroid · EyePACS k=10 | accuracy: 0.611 -> 0.705 | Improved |
| add local style alignment instead of Euclidean averaging | k-Centroid · APTOS k=2 | accuracy: 0.688 -> 0.687 | No clear effect |
| add local style alignment instead of Euclidean averaging | k-Centroid · APTOS k=10 | MIA top-K accuracy (%): 2.1 -> 1.0 | Improved |
| use GAN-based averaging instead of pixel averaging | k-Same-Pixel · APTOS k=5 | FID: 131.69 -> 20.09 | Improved |
| add local style loss to content loss (ablation) | k-SALSA · APTOS k=5 | accuracy: 0.673 -> 0.712 | Improved |
| add content loss to local style loss (ablation) | k-SALSA · APTOS k=5 | accuracy: 0.685 -> 0.712 | Improved |
| use local instead of global style features (ablation) | k-SALSA · APTOS k=5 | Cohen's kappa: 0.761 -> 0.769 | Improved |
| add flexible alignment to local style loss (ablation) | k-SALSA · APTOS k=5 | accuracy: 0.57 -> 0.712 | Improved |
| add flexible alignment to local style loss (ablation) | k-SALSA · APTOS k=5 | Cohen's kappa: 0.417 -> 0.769 | Improved |
Closest papers by meaning
- FOCUS: Benchmarking Retinal Model Generalization from Foundation Vision Encoders to Multimodal LLMs (NeurIPS 2026 · similarity 0.88)
- Domain Adaptive Diabetic Retinopathy Grading with Model Absence and Flowing Data (CVPR 2025 · similarity 0.87)
- Pathological Evidence Exploration in Deep Retinal Image Diagnosis (AAAI 2019 · similarity 0.85)
- Like an Ophthalmologist: Dynamic Selection Driven Multi-View Learning for Diabetic Retinopathy Grading (AAAI 2025 · similarity 0.84)
- SIGraph: Saliency Image-Graph Network for Retinal Disease Classification in Fundus Image (AAAI 2025 · similarity 0.83)
- Direct Discriminative Optimization: Your Likelihood-Based Visual Generative Model is Secretly a GAN Discriminator (ICML 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.
