People gatherings, including political rallies, religious gatherings, and conferences, are vulnerable to security threats from individuals with criminal backgrounds. Even if a bad person gets into such a situation, the consequences can be devastating. Current security measures often rely on manual controls and are prone to human error. They lack the ability to quickly and legally identify people with criminal histories, resulting in a cumulative effect on crime. This paper aims to improve security by using computerized access to meetings to identify and supervise participants. The system will use facial recognition technology to identify people through video surveillance and still images and various predefined libraries and tools, including scikit-learn, dlib, Open CV, Tensor Flow, facial recognition library and Py-Charm. Utilizing existing face detection and feature extraction techniques, we introduce a methodology EPIANN (Efficient Person Identification using Artificial Neural Networks) that assesses our approach based on evaluation metrics, including precision, recall, and F1-score. Based on the research results, clearly indicates that the EPIANN model integrated with OpenCV for real-time person recognition within every frame offers the highest performance, achieving a recall: 0.93%, precision: 0.91%, and F1-score: 0.920%. Ultimately, it can be confidently asserted that the EPIANN approach surpasses alternative methodologies.