Computer Vision

Course dates – HT – week beginning Monday 14th January 2019 – for Year 1 students
Andrew Zisserman, Andrea Vedaldi and Victor Prisacariu

Introduction

Computer vision is empowering cutting-edge applications in search, smart sensing, medical imaging, human-machine interaction, and many other areas. In this course the students will be introduced to the fundamental theory and practice of this rapidly evolving technology and will learn the fundamentals required to make use of it in their own research projects.

Objectives
  • To introduce state-of-the-art methods for computer vision
  • To familiarise the student with the theory and practice of image matching and indexing
  • To reveal the geometry underpinning the formation of images from observations of the 3D world
  • To supply computer vision software that can be used in subsequent research
Contents
  • Object recognition
    • Image transformations and matching
    • Image indexing and search
    • Sliding-window object detectors
  • Multi-view geometry
    • Camera models
    • Image correspondences
    • Triangulation
    • 3D reconstruction of motion and structure
  • Differential motion
    • optical flow
    • object tracking
  • Segmentation
    • edges
    • superpixels
Prerequisites
  • MATLAB
  • Basic linear algebra
Other Sources
  • Computer Vision: A Modern Approach by David Forsyth and Jean Ponce (2nd ed.)
  • Computer Vision: Algorithms and Applications by Richard Szeliski (PDF available online)
  • Multiple View Geometry in Computer Vision by Richard Hartley and Andrew Zisserman
Exercises
  • Practical on image matching and indexing
  • Practical on sliding window object detection
  • Practical on multi-view geometry
Assessment Mode
  • Continual assessment during practicals