"Multi-Class 13-Lead Paper ECG Image Dataset with 1D Signal Reconstruction"

"The Multi-Class 13-Lead Paper ECG Image Dataset with 1D Signal Reconstruction is developed by the research and engineering team at Pixemantic to support advancements in medical image processing, ECG digitization, and AI-driven cardiovascular diagnosis.This dataset contains 1,288 digitized paper-based ECG records categorized into five clinically relevant classes: Normal (250), Myocardial Infarction (240), History of Myocardial Infarction (166), Abnormal Heartbeat\/Arrhythmia (233), and Other Abnormal patterns (399).Each record includes the original ECG paper image, 13 segmented lead images, and reconstructed 1D digital signals extracted using image processing, thresholding, smoothing, and waveform recovery techniques.The dataset enables lead-wise signal analysis, extraction of P, QRS, and T wave components, and multi-class classification research. It is intended to facilitate the development of automated ECG interpretation systems, deep learning models, and explainable AI solutions for cardiac abnormality detection and legacy paper ECG digitization."