Download this file

22 lines (12 with data), 905 Bytes

ECG UNET

Re-production of the model proposed in Reference [1] using LUDB

Training

A typical experiment of training a U-Net model for wave delineation

Curves of train loss and validation f1 score Curves of the mean error and standard deviation
score-loss me-std

Evaluation

An example of a trained wave delineation U-Net model evaluated on a 10s segment from the validation set

val-example

References

  1. Moskalenko, Viktor, Nikolai Zolotykh, and Grigory Osipov. "Deep Learning for ECG Segmentation." International Conference on Neuroinformatics. Springer, Cham, 2019.