SkinDeepRESEARCHSteve Seguin

Use case · Idea to test

Check ECG recording quality

Flag an unusable trace for another recording.

  1. Recorded heart trace
  2. Check signal quality
  3. Accept or request review
Illustration of the proposed workflow.

A recording tool could flag poor signal quality before someone relies on the trace. This is a different task from deciding whether the heart rhythm is abnormal.

Where the small model fits

A small signal classifier returns a quality label. An adaptive model could spend more processing on unclear recordings, or leave the quality decision for review instead of forcing an answer.

What would need to work

Compare with an ordinary signal-quality classifier, including time spent on exit checks. Test bad recordings accepted and good ones rejected across devices. A quality label is not a normal-heart verdict. This lab has not run an ECG quality experiment.

Research proposal only. SkinDeep has no clinical model, patient-use demo or validated medical performance for this task.

A dataset to start with

The PhysioNet 2011 challenge provides an ECG recording-quality task, separate from disease classification. PhysioNet ECG-quality challenge.