Face recognition using deep learning is a highly active research area in computer vision that focuses on identifying or verifying individuals from facial images or videos with high accuracy. Early deep learning approaches leveraged convolutional neural networks (CNNs) to learn robust facial feature representations, while landmark-based preprocessing and alignment techniques improved recognition performance. Subsequent research introduced deep metric learning approaches such as DeepFace, FaceNet, and SphereFace, which use embedding spaces with triplet or angular loss functions to enhance discriminability and generalization. Recent advances explore attention mechanisms, residual networks, transformer-based architectures, and domain adaptation techniques to handle pose, illumination, occlusion, and aging variations. Applications span security and surveillance, authentication systems, social media, human–computer interaction, and forensics. Current studies also focus on lightweight and privacy-preserving face recognition models suitable for edge devices, adversarial robustness, and multi-modal fusion with infrared or depth data to improve reliability and deployment in real-world scenarios.