This repository implements MTCNN (Multi-Task Cascaded Convolutional Neural Networks) for face detection and alignment. The system uses a three-stage cascaded architecture to progressively refine face detection and predict five facial keypoints. It runs at real-time speeds (>30 FPS) on CPU and supports deployment across Windows, Ubuntu, Mac, Android, and iOS using frameworks like Caffe, TensorFlow, and NCNN.