Research map / Adversarial robustness and security / Certified and empirical robustness
Toward Robust Diagnosis: A Contour Attention Preserving Adversarial Defense for COVID-19 Detection
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use CAP instead of FAT | Visformer-Tiny · SARS-COV-2 | average adversarial accuracy (%): 82.62 -> 86.96 | Helped |
| Use CAP instead of TRADES | Visformer-Tiny · COVID19-C | average OoD adversarial accuracy (%): 46.60 -> 49.56 | Helped |
| Use CAP instead of AVmixup | Deit-Tiny · SARS-COV-2 | average adversarial accuracy (%): 84.79 -> 86.28 | Helped |
| Use CAP instead of AT | Visformer-Tiny · MosMed-L | total adversarial accuracy: 80.9 -> 85.6 | Helped |
| Use CAP instead of TRADES | Visformer-Tiny · MosMed-L to MosMed-M | total clean/adversarial accuracy: 26.2/14.5 -> 61.2/58.7 | Helped |
| Add self-guided filtering (SGF) (ablation) | TRADES, Visformer-Tiny · SARS-COV-2 | PGD-10 accuracy (%): 77.67 -> 79.68 | Helped |
| Add self-guided filtering (SGF) (ablation) | TRADES, Visformer-Tiny · COV-M | accuracy (%): 47.38 -> 50.11 | Helped |
| Add attention parameter regularization (AWR) (ablation) | CAP with SGF, Visformer-Tiny · SARS-COV-2 | PGD-10 accuracy (%): 79.68 -> 83.50 | Helped |
| Add attention parameter regularization (AWR) (ablation) | CAP with SGF, Visformer-Tiny · COV-M | accuracy (%): 50.11 -> 51.71 | Helped |
| Add hybrid distance metric (HDM) (ablation) | CAP with SGF and AWR, Visformer-Tiny · SARS-COV-2 | PGD-10 accuracy (%): 83.50 -> 85.92 | Helped |
| Add hybrid distance metric (HDM) (ablation) | CAP with SGF and AWR, Visformer-Tiny · COV-M | accuracy (%): 51.71 -> 53.08 | Helped |
| Use SE for outer and inner optimization instead of outer SE and inner KL (ablation) | CAP, Visformer-Tiny · SARS-COV-2 | PGD-10 accuracy (%): 85.92 -> 63.68 | Hurt |
Closest papers by meaning
- Large-scale and Fine-grained Vision-language Pre-training for Enhanced CT Image Understanding (ICLR 2025 · similarity 0.82)
- NaFV-Net: An Adversarial Four-view Network for Mammogram Classification (AAAI 2025 · similarity 0.82)
- Cross-Modal Interactive Perception Network with Mamba for Lung Tumor Segmentation in PET-CT Images (CVPR 2025 · similarity 0.81)
- Domain Generalized Medical Landmark Detection via Robust Boundary-Aware Pre-Training (AAAI 2025 · similarity 0.81)
- Attacking CNNs in Histopathology with SNAP: Sporadic and Naturalistic Adversarial Patches (Student Abstract) (AAAI 2024 · similarity 0.81)
- ConDSeg: A General Medical Image Segmentation Framework via Contrast-Driven Feature Enhancement (AAAI 2025 · similarity 0.81)
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.