Facial verification is a critical application in various industries such as security, law enforcement, and access control systems. In this project, we propose a face verification system that combines state-of-the-art technologies, including YOLO (You Only Look Once) for object detection, FaceNet for facial feature extraction, and ArcFace for model training. The system utilizes YOLO to locate and extract faces from input images, and then FaceNet to extract high-quality facial features. We then use ArcFace to train our model for improved accuracy and robustness. The system calculates the similarity between faces using Euclidean distance, a widely used metric for measuring distance in high-dimensional spaces. The proposed methodology involves collecting a large dataset of images containing faces, preprocessing the data, training our model, and evaluating its performance using various metrics such as accuracy, precision, and recall. By following this methodology, we can build a robust and accurate face verification system that can be deployed in real-world scenarios to improve efficiency and effectiveness.