In recent years, important breakthroughs have been made in high-level computer vision, including object, scene, and activity recognition. These accomplishments have been made feasible by recent advances in machine learning, statistical image modeling, sparse coding, graph/information theory, and other methodologies. The course surveys the latest trends in high-level vision in terms of new problem and mathematical-theory formulations.
- (40%) Oral presentations
- (50%) Final project
- (10%) In-class discussions
- Develop breadth and depth in understanding of, and critical thinking to the state of the art
- Develop presentation skills to communicate clearly, precisely, and concisely complex concepts
- Develop writing skills to systematize and synthesize other people's ideas
This course is aimed at graduate students who already have some knowledge of probability and statistics, machine learning, artificial intelligence, and linear algebra.