ECG IMAGES

This dataset is composed of Electrocardiogram (ECG) images obtained from the database MIT-BIH Arrhythmia. For that, the ECG signals were pre-processed, generating 109.445 images with a resolution of 256x256. In sequence, five cardiac arrhythmia superclasses recommended by AAMI were selected for work.

ECG images dataset:

Number of Samples: 109445
Number of Categories: 5
Image Resolution: 256x256
Data Source: Physionet's MIT-BIH Arrhythmia Dataset
Classes: [N, S, V, F, Q]

More information about the acquisition of this dataset can be found at: https://github.com/analiviafr/arrhythmia_classifier