Quantified Explainability: Convolutional Neural Network Focus Assessment in Arrhythmia Detection

This dataset consists of electrocardiographic (ECG) images, each containing 6 heartbeats, converted from the MIT BIH arrhythmia database. Only the MLII/V5 leads were converted. Each folder is identified by the participant ID and the figures of each folder correspond to the specific ECG records. Additionally, the labels of each figure are present in a .txt file inside each folder.On the other hand, the .pth files consist of CNN (Resnet-50) trained on the records 101, 106, 109, 112, 115, 116, 118, 119, 122, 124, 203, 205, 208, 209, 220, 223 of the dataset.