Learning-based Vision

Course Description

This advanced undergraduate/graduate-level course offers an in-depth exploration of modern computer vision techniques centered on deep learning. While traditional computer vision focuses on geometry and handcrafted features, this course covers the end-to-end learning paradigms that have revolutionized the field. We will cover the foundations of image recognition, including the design and training of convolutional neural networks (CNNs), and progress to advanced, state-of-the-art topics. Key areas of study will include self-supervised learning, Vision Transformers (ViTs), generative modeling (VAEs, GANs, and diffusion models), neural rendering with Neural Radiance Fields (NeRFs), and the rise of large-scale, prompt-guided vision models. The course will emphasize both the theoretical underpinnings and the practical implementation of these models. Students will gain hands-on experience through programming assignments and a final project, preparing them for research or advanced development roles in computer vision. Recommended background: computer vision, machine learning and deep learning. Prerequisite(s): Students may earn credit for EN.601.692 OR 492, not both.;( EN.601.461 OR EN.601.661 ) OR ( EN.601.482 OR EN.601.682 )

Code

EN.601.492

Credits

3

Subject

Computer Science