Computer Vision

MSc - Autumn Semester
102470, Lectures and exercises, 5.0 ECTS

Lecturer Prof. Dr. Paolo Favaro
Teaching assistants Mr. Adrian Wälchli
Mr. Xiaochen Wang
Mr. Adam Bielski
Location Seminarraum 002, Engehaldenstrasse 8
Time Tuesdays 14.15-16.00 (lecture), and 16.15-17.00 (exercises)
Exam February 11 2020 from 10:00 to 12:00 at Seminarraum 002, Engehaldenstrasse 8
ILIAS KSL

Course description

This course covers fundamental topics in computer vision. The course will provide an introduction to image formation, image processing, feature detection, segmentation, multiple view geometry and 3D reconstruction, and motion.

Learning outcome

Upon successful completion of this class, you will be able to:

  1. Understand how cameras capture images of a scene Implement
  2. Implement and use: algorithms for image processing such as image filtering and image segmentation;
    algorithms for object detection (such as faces) and recognition;
    algorithms for 3D reconstruction (e.g., from stereo systems) 
  3. Describe the mathematics underpinning each method and know how to adapt it to new scenarios.

Prerequisites

The course requires students to be familiar with the basics of linear algebra, probability theory and Python programming. A brief review of these subjects will be carried out during the exercise sessions.

Resources

The following books are recommended as additional reading:

  • Computer Vision : A Modern Approach, David A. Forsyth and Jean Ponce.
  • Pattern Recognition and Machine Learning, Christopher Bishop.
  • Algorithms and Applications, Rick Szeliski.
    An electronic copy is available for free online
  • Visual Object Recognition, Kristen Grauman and Bastian Leibe.
    This book is also available online for free.
  • Computer Vision: Models, Learning, and Inference, Simon J.D. Prince.

Assessment

The exercises are a prerequisite for registering for the exam. There will be several homework assignments and the deadlines will be specified by the teaching assistant. For admission to the examination one must pass every assignment. The assignments will also contribute to 30% of the final mark. The description of each assignment will be made available in ILIAS. Notice that many assignments will require use of Python. All the homework assignments are also intended for exam preparation.

 

Schedule and material

The following table provides an overview of the content of the lectures during the semester. Please check it periodically as it might be updated.

Week Lecture Reading
1 Introduction & projection models Forsyth & Ponce Ch 1
2 Camera models & Linear filters & edges Forsyth & Ponce Ch 1, 2, 7, 8
3 Energy minimization methods
4 Photometric stereo & shading Forsyth & Ponce Ch 5
5 Tracking & Optical flow Forsyth & Ponce Ch 17
6 Interest points detection
7 Registration & Fitting Forsyth & Ponce Ch 15, 17
8 Epipolar geometry & Stereo Forsyth & Ponce Ch 10, 15
9 Multiview stereo & Structure from motion Forsyth & Ponce Ch 11, 12, 13
10 Recognition & machine learning Grauman & Leibe
11 Bayesian methods Forsyth & Ponce Ch 16, C. Bishop, S. Prince
12 Bayesian methods
13 Segmentation
14 Revision Handouts