CSC-486 Deep Learning in Vision (3)

This course provides an overview of deep neural network, also known as deep learning, and related methods with applications in computer vision, particularly in image classification. It reviews fundamental topics in computer vision including lineaging filtering, sampling, Fourier transform, texture analysis, 3D geometry, and camera models. It introduces machine learning methods and neural network architectures. The course also dives into a variety of deep learning-based applications in computer vision including image classification, representational learning, image segmentation, and image synthesis. Students learn to implement, train, and debug off-the-shelf neural networks and gain a detailed understanding of the cutting-edge research in computer vision. Subjects include visual data representation, mid-level vision, image parsing, image classification and synthesis with architectures such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), attention and transformers, and generative models (GANs, diffusion models). Prerequisite: CSC-476 and MATH-310.