teaching learning-based methods for computer vision (UK CS 585/685)

01 January 2015

Course Summary

This course covers the use of methods from machine learning to interpret imagery. The focus will be on reading recent and historical papers.

Pre-requisites

Basic knowledge of machine learning (equivalent to CS 460g). Some experience with image processing or computer vision will be helpful, but is not required. Please ask the instructor if you are unsure if you have sufficient experience.

Student Learning Outcomes

After completing this course, the student will be able to:

• Describe the common problem definitions in image and video interpretation.
• Describe the fundamental challenges in solving each of these problems.
• Describe several approaches to solving these problems and how these address the fundamental challenges.
• Determine the capabilities and limitations of a learning-based method for computer vision.
• Compare and contrast approaches to solving image and video interpretation problems.
• Evaluate the strengths and weaknesses of an academic paper on learning-based methods for computer vision.
• Describe recent developments in learning-based methods for computer vision.
• Create and describe a novel learning-based approach to solving an image interpretation problem.