This is code for ECG segmentaion using 1 dimensional U-net
I refered the paper "Deep Learning for ECG Segmentation"[paper]
Lobachevsky University Electrocardiography Database (LUDB) was used for ECG segmentaion task. link
It contains 12 lead ECG data from 200 patients for 10 senconds. I preprocessed and extracted ECG segments and its code was included in notebook.
In data folder, preprocessed files were contained. Rule of file name is "{ptient_id}_{segment number}.npz" I split train and test data by patients id.
Segments model was one dimensional U-net (1D U-net). Model structure was same with the model suggested by the paper.
The results of ECG segmenation
Ajou University, Republic of Korea
Computational Medical Intelligence (CMI) Lab
E-mail : jangood1122@gmail.com