The Advanced Computer Vision and Deep Learning course on Udacity covers a range of topics in advanced CNN architectures. Students will learn about region-based CNN's such as Fast R-CNN and Faster-R-CNN, which enable fast and localized object recognition in images. The course also covers YOLO, a real-time object detection algorithm, and explores concepts like grid, sliding windows, and bounding boxes in object detection. Students will gain an understanding of recurrent neural networks (RNNs) and their applications in deep learning, as well as how feedforward and backpropagation through time work. The course also delves into Long Short-Term Memory Networks (LSTMs) and how memory can be incorporated into deep learning models. Important hyperparameters like learning rate, epochs, and layers are refreshed, with a focus on hyperparameters in RNNs. Optional topics include attention mechanisms and their applications in natural language and computer vision, as well as basic attention methods like additive attention, Bahdanau, and Luong attention. The course concludes with a project on image captioning, where students combine their knowledge of CNNs and RNNs to build a deep learning model that generates captions for input images.