Reproduce “BeatFlow-ECG: Rectified Flow for ECG Reconstruction from Indirect Wearable Signals” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Mohamed Kamel, Sahar Selim, Walaa Medhat, Tamer Nadeem
BeatFlow-ECG reconstructs single-channel ECG waveforms from wrist PPG and inertial measurements using rectified flow, iteratively updating a noise-based waveform conditioned on synchronized signals and motion. Under leave-one-subject-out evaluation on PPG-DaLiA, it outperformed Conditional DDPM-1D, with mean L1 error 0.062 versus 0.085 and mean R-peak F1 0.946 versus 0.905. The evidence covers two datasets and does not establish diagnostic validity.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.