Biometric authentication has emerged as a cornerstone of secure identity verification systems. Among the diverse modalities, electrocardiogram (ECG)-based authentication stands out for its physiological uniqueness and integration potential with wearable devices. This study introduces an optimized convolutional neural network (CNN)-based framework for ECG-based authentication, achieving a high accuracy of 99.62% with minimal computational overhead. The study outlines which combine structure-based progresses and innovative preprocessing approaches to develop ECG-based biometric authentication, so as to be robust and reliable for practical healthcare-based applications. The proposed method makes sure that the system should manage the encounters of varied and noisy environments which come across in biomedical contexts. The presented work offers an inclusive comparison of ECG-based biometrics with traditional methods like fingerprint and iris recognition. The results highlight that ECG signals not only deliver higher security and robustness but are also highly adaptable. ECG data captures dynamic physiological processes, making it less vulnerable to spoofing or environmental disruptions. The study also addresses challenges in implementing ECG-based systems in real-world applications. Issues like data variability caused by individual health conditions, hardware constraints, and signal noise are thoroughly discussed. A more robust potential solution is proposed, with advanced signal processing, adaptive learning models, and the integration of ECG with other biometric methods to develop multimodal systems. These multimodal approaches that could combine ECG with other biometrics to improve accuracy and reliability while broadening application possibilities. This research further discovers future opportunities for scaling ECG-based systems. The study emphasizes the potential of ECG as a key technology for secure, non-invasive biometric authentication, paving the way for its wider adoption in health care and other sectors.