Physiologicalsignal analysis
Beat-accurate R-peak detection and rhythm analysis for multi-channel ECG, built to sit inside your evaluation software.
Concept illustration – not real results
Built for teams that make ECG software
Annotating long ECG recordings by hand is slow, and the beat detection bundled with many evaluation tools leaves reviewers correcting more than they confirm.
Our pipeline does the first pass: it finds every beat, labels it and measures the intervals in between, so clinical reviewers start from an annotation worth checking.
R-peak detection
Locates each beat across noisy multi-lead recordings. Benchmarked against Pan-Tompkins and NeuroKit2 on the MIT-BIH Arrhythmia Database; results are published here once final.
Beat and rhythm analysis
Labels beats and flags premature or irregular rhythm, with the RR series and its compensatory pauses made explicit.
Pre-annotation for review
Annotations are exported in the format your evaluation software reads, so a nurse or physician confirms rather than starts from scratch.
Synthetic ECG for illustration – no patient data, not product output.
Specification
- Input
- EDF / EDF+, raw sample arrays
- Channels
- 1–3 leads
- Output
- R-peak indices, beat labels, RR series
- Delivery
- OEM SDK licence
- Validation
- MIT-BIH Arrhythmia Database (in progress)
- Documentation
- IEC 62304 lifecycle (planned)
Where it stands
- Prototype pipelineWorking end to end on multi-lead recordings
- Benchmark against Pan-Tompkins and NeuroKit2In progress on MIT-BIH
- SDK packaging and IEC 62304 documentationPlanned
Licensing and pilot projects
If your software reads ECG recordings, we can run a pilot on a sample of your data and show the result side by side with your current annotation.