Resting ECG Segmentation Dataset

Resting ECG Segmentation Dataset (RDB) is a fully annotated, 12-lead resting-state ECG corpus built for waveform‐level delineation research.Origin – RDB is a derivative of the large, open ChapmanECG arrhythmia database (Zheng et al., 2020).Scope – 2 399 recordings, each 10 s long, sampled at 500 Hz.Annotations – Every record is labelled for P-waves, QRS complexes and T-waves on all 12 leads. Initial marks were placed by trained physicians and subsequently reviewed and refined by senior cardiologists; each lead was annotated independently to preserve lead-specific morphology.Rhythm diversity – The set spans atrial fibrillation (AFIB), atrial flutter (AF), atrial tachycardia (AT), supraventricular tachycardia (SVT) and multiple sinus irregularities, providing rich morphological variation for robust model training.Rhythm Type Records numAF (Atrial Flutter) 400AFIB (Atrial Fibrillation) 400AT (Atrial Tachycardia) 121SB (Sinus Bradycardia) 400SI (Sinus Irregularity) 399SR (Sinus Rhythm) 400ST (Sinus Tachycardia) 140SVT (Supraventricular Tachycardia) 139Total 2399The combination of high-quality beat-level labels and broad rhythm coverage makes RDB a strong benchmark for developing and evaluating ECG segmentation algorithms that must generalise across diverse clinical presentations.DOI:10.6084/m9.figshare.28892186Citations: Zheng, J., Zhang, J., Danioko, S. et al. A 12-lead electrocardiogram database for arrhythmia research covering more than 10,000 patients. Sci Data 7, 48 (2020). DOI: https://doi.org/10.1038/s41597-020-0386-x.Liu, Y., Zhang, P., Feng, X., Hu, D., Zhou, D., Li, J., ... & Li, Q. (2025). Y-Net-ECG: A Multi-Lead informed and interpretable architecture for ECG segmentation across diverse rhythms. Expert Systems with Applications, 127955. DOI: https://doi.org/10.1016/j.eswa.2025.127955.