Convolutional Neural Networks (Andrew Ng, DeepLearning.AI)

Course 4 of the Deep Learning Specialization in Andrew Ng's lecture videos: 42 lessons, about 6 hours, on computer vision. Week 1 builds convolutions from edge detection, padding, strides and convolutions over volumes to pooling layers and a full CNN example. Week 2 studies classic networks, ResNets, network-in-network and Inception, plus transfer learning and data augmentation. Week 3 covers object detection: localisation, landmarks, intersection over union, non-max suppression, anchor boxes, YOLO and region proposals. Week 4 covers face recognition with one-shot learning, Siamese networks and triplet loss, then neural style transfer. A solid foundation for image-based AI projects.