Deep Learning for ECG Segmentation
This work addresses the problem of accurate and efficient ECG segmentation for medical diagnostics, representing an incremental improvement with strong specific gains.
The authors tackled ECG segmentation by proposing a UNet-like convolutional neural network that adapts to different sampling rates and ECG monitor types, achieving F1-measures of at least 97.8% for P waves, 99.5% for T waves, and 99.9% for QRS complexes.
We propose an algorithm for electrocardiogram (ECG) segmentation using a UNet-like full-convolutional neural network. The algorithm receives an arbitrary sampling rate ECG signal as an input, and gives a list of onsets and offsets of P and T waves and QRS complexes as output. Our method of segmentation differs from others in speed, a small number of parameters and a good generalization: it is adaptive to different sampling rates and it is generalized to various types of ECG monitors. The proposed approach is superior to other state-of-the-art segmentation methods in terms of quality. In particular, F1-measures for detection of onsets and offsets of P and T waves and for QRS-complexes are at least 97.8%, 99.5%, and 99.9%, respectively.