ECGWmark: Deep Watermarking for Electrocardiogram Signals
ECGWmark watermarks ECG signals without quantization and keeps over 90% bit accuracy under circular shifts.
ECGWmark is a fully convolutional, quantization-free deep watermark for electrocardiogram signals intended to carry patient metadata in telecardiology. It works in a continuous full-resolution latent space, uses frame-level embedding against temporal desynchronization, and applies a differentiable physiological loss guided by a clinical classifier. On MIT-BIH, channel-bit extraction stays above 90% under evaluated circular shifts, while inter-patient diagnostics change unevenly: Class-V sensitivity moves from 71.97% to 76.50% and Class-S sensitivity from 8.89% to 1.39%.
- Quantization-free convolutional watermark targets ECG time-series metadata.
- Bit extraction stays above 90% under tested circular shifts on MIT-BIH.
- Inter-patient diagnostic metrics show mixed class-level effects.
From the original article
1 min read at arxiv.orgSecure transmission of electrocardiogram (ECG) signals remains a major challenge in telecardiology. While traditional mathematical watermarking can embed patient metadata directly into ECG signals, it is highly fragile against temporal desynchronization such as circular phase shifts.
By Kuan-Yu Chen · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.