WPW Detector
Upload a 12-lead ECG. The detectors return a suspicion level from 1 to 100, and show you what pushed it.
An experimental research tool. Not a validated medical device, and not a substitute for a physician.
Before you begin
- This is an experimental research tool, not a validated medical device.
- It does not provide a diagnosis and does not replace a doctor.
- The result is a suspicion rank, not a probability or a certainty.
- It can be wrong in both directions: it may miss a real case, or flag a normal ECG.
- If you have symptoms or any concern, consult a physician.
Your data, and your consent
Your ECG is sensitive health information. Please confirm the following before continuing.
Upload your ECG
Add your 12-lead ECG recording. The file stays on the server only long enough to be read.
Pick a detector
The ensemble is recommended. Each detector shows its average precision (AP).AP (average precision) is the metric here because WPW is very rare, so plain accuracy would be misleading: a detector that always answered "no" would be right 99.8% of the time. These figures come from a held-out test set that was never used during development. The confidence intervals are wide because that test set holds very few WPW cases, so small differences between detectors are not meaningful.
Analyzing
Reading the signal, beat by beat.
These are the signal patterns that pushed the reading up or down. Careful: they explain what the model saw in the signal, not the score above. Because WPW is so rare, even a faint signal is enough to raise the score, which is why the patterns can lean against suspicion while the score is still not zero.