Journal of Artificial Intelligence and Metaheuristics JAIM 2833-5597 10.54216/JAIM https://www.americaspg.com/journals/show/2111 2022 2022 South Ural State University Department of food and biotechnology Chelyabinsk, Russia Ammar Kadi Laboratoire de Technologie Industrielle et de l’Information, Faculté de Technologie, Université de Bejaia, 06000 Bejaia, Algeria Adel Oubelaid Computer Science and Intelligent Systems Research Center, Blacksburg 24060, Virginia, USA S. K. Towfek The electrocardiogram (ECG) is a type of biometric data that has recently attracted a lot of attention as a potentially useful biometric trait due to its high discriminatory power. However, precise and consistent biometric identification systems are challenging to deploy because to ECG signals' vulnerability to a wide range of sounds, including baseline wander, powerline interference, and highlow frequency noises. That's why ECG signal denoising is such an important aspect of the preprocessing phase for ECG-based biometric person identification: it removes noise from the raw ECG data. Biometric recognition using ECG signals is a difficult problem involving phases of preprocessing, feature extraction, feature selection, feature modification, and classification. Biometric system analysis also relies heavily on the use of appropriate success measures and a well-organized library of ECG signals. This is especially crucial when considering the fact that researchers rely significantly on freely accessible resources to gauge the efficacy of the algorithms they propose. In this study, we examine most of the approaches that have been taken toward ECG-based biometric verification of humans. 2023 2023 31 40 10.54216/JAIM.050203 https://www.americaspg.com/articleinfo/28/show/2111