Synthetic ECG signal generation, a review

An electrocardiogram (ECG/EKG) records the heart's electrical activity and is used to diagnose cardiovascular disease and heart health. Therefore, valuable information about the patient's cardiac condition can be obtained to prevent heart attacks as much as possible. Diagnosing these cardiac diseases requires sufficient data for each category. However, in today's databases, data related to irregular ECG signals are much rarer than regular heartbeats, leading to imbalanced datasets and low accuracy in classifying minority categories. Additionally, patient privacy is also a concern. Considering these two issues, generating artificial ECG signals is essential to balance the dataset and preserve patient privacy. Recently, several techniques have been proposed for generating artificial ECG signals. This article aims to provide an overview of ECG signal generation methods and their related aspects, classify research in this field, and discuss important points and challenges associated with these methods. The evaluation metrics used to assess the generated ECG signals are also explained.