{"ID":2827025,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.17423","arxiv_id":"2512.17423","title":"SCAR: Semantic Cardiac Adversarial Representation via Spatiotemporal Manifold Optimization in ECG","abstract":"Deep learning models for Electrocardiogram (ECG) analysis have achieved expert-level performance but remain vulnerable to adversarial attacks. However, applying Universal Adversarial Perturbations (UAP) to ECG signals presents a unique challenge: standard imperceptible noise constraints (e.g., 10 uV) fail to generate effective universal attacks due to the high inter-subject variability of cardiac waveforms. Furthermore, traditional \"invisible\" attacks are easily dismissed by clinicians as technical artifacts, failing to compromise the human-in-the-loop diagnostic pipeline. In this study, we propose SCAR (Semantic Cardiac Adversarial Representation), a novel UAP framework tailored to bypass the clinical \"Human Firewall.\" Unlike traditional approaches, SCAR integrates spatiotemporal smoothing (W=25, approx. 50ms), spectral consistency (\u003c15 Hz), and anatomical amplitude constraints (\u003c0.2 mV) directly into the gradient optimization manifold. Results: We benchmarked SCAR against a rigorous baseline (Standard Universal DeepFool with post-hoc physiological filtering). While the baseline suffers a performance collapse (~16% success rate on transfer tasks), SCAR maintains robust transferability (58.09% on ResNet) and achieves 82.46% success on the source model. Crucially, clinical analysis reveals an emergent targeted behavior: SCAR specifically converges to forging Myocardial Infarction features (90.2% misdiagnosis) by mathematically reconstructing pathological ST-segment elevations. Finally, we demonstrate that SCAR serves a dual purpose: it not only functions as a robust data augmentation strategy for Hybrid Adversarial Training, offering optimal clinical defense, but also provides effective educational samples for training clinicians to recognize low-cost, AI-targeted semantic forgeries.","short_abstract":"Deep learning models for Electrocardiogram (ECG) analysis have achieved expert-level performance but remain vulnerable to adversarial attacks. However, applying Universal Adversarial Perturbations (UAP) to ECG signals presents a unique challenge: standard imperceptible noise constraints (e.g., 10 uV) fail to generate e...","url_abs":"https://arxiv.org/abs/2512.17423","url_pdf":"https://arxiv.org/pdf/2512.17423v1","authors":"[\"Shunbo Jia\",\"Caizhi Liao\"]","published":"2025-12-19T10:22:53Z","proceeding":"eess.SP","tasks":"[\"eess.SP\"]","methods":"[]","has_code":false}