Page 45 -

24 1 Introduction

Week Material Project
(1.) Chapter 2 Image formation
2. Chapter 3 Image processing
3. Chapter 4 Feature detection and matching P1
4. Chapter 6 Feature-based alignment
5. Chapter 9 Image stitching P2
6. Chapter 8 Dense motion estimation
7. Chapter 7 Structure from motion PP
8. Chapter 14 Recognition
(9.) Chapter 10 Computational photography
10. Chapter 11 Stereo correspondence
(11.) Chapter 12 3D reconstruction
12. Chapter 13 Image-based rendering
13. Final project presentations FP
Table 1.1 Sample syllabi for 10-week and 13-week courses. The weeks in parentheses are not used in the shorter
version. P1 and P2 are two early-term mini-projects, PP is when the (student-selected) final project proposals are
due, and FP is the final project presentations.

or related mathematical techniques. (See also the Introduction to Computer Vision course at
Stanford, 12 which uses a similar curriculum.) Related courses have also been taught on the
topics of 3D photography 13 and computational photography. 14
When Steve and I teach the course, we prefer to give the students several small program-
ming projects early in the course rather than focusing on written homework or quizzes. With
a suitable choice of topics, it is possible for these projects to build on each other. For exam-
ple, introducing feature matching early on can be used in a second assignment to do image
alignment and stitching. Alternatively, direct (optical flow) techniques can be used to do the
alignment and more focus can be put on either graph cut seam selection or multi-resolution
blending techniques.
We also ask the students to propose a final project (we provide a set of suggested topics
for those who need ideas) by the middle of the course and reserve the last week of the class
for student presentations. With any luck, some of these final projects can actually turn into
conference submissions!
No matter how you decide to structure the course or how you choose to use this book, I
encourage you to try at least a few small programming tasks to get a good feel for how vision
techniques work, and when they do not. Better yet, pick topics that are fun and can be used on
your own photographs, and try to push your creative boundaries to come up with surprising
results.

12 http://vision.stanford.edu/teaching/cs223b/
13 http://www.cs.washington.edu/education/courses/558/06sp/
14 http://graphics.cs.cmu.edu/courses/15-463/