Syllabus

The recommended sequence for studying each topic is as follows:

  • Watch the lecture video(s).
  • Read and study the class notes.
  • Attend and engage in the discussion section.
  • Do the homework if published.

Zoom participation links are on the course home page.

Links to videos, lecture notes, slides used in the videos, and recordings of discussion sessions will show up in the table below as the semester progresses. Protected links to Warpwire libraries with all the recordings are also available as follows:

The slides used in the videos are provided only for convenience below. They are not required reading. Slides should not be used as your main study materials. That is what the class notes and videos are for.

Please see the mechanics page for details about discussions and homework.

There is nothing for you to prepare or any video to watch for the first discussion session, which is on January 21st. However, you may want to get a head start on homework 0, on prerequisites, available on the homework page.

Week Topic Notes Slides Videos Discussion HW Due
0 Course Content and Logistics s_0 d_00_1, d_00_2 Jan 28
1 Correlation, Convolution, Filtering n_01 s_01 v_01 d_01_1, d_01_2 Feb 4
2 Image Differentiation and Pyramids n_02 s_02 v_02 d_02_1, d_02_2 Feb 15
3 Numerical Function Optimization n_03 s_03 v_03 d_03_1, d_03_2
4 Deep Learning Basics n_04 s_04 v_04 d_04_1, d_04_2 Feb 25
5 Backprop, Networks for Recognition n_05 s_05 v_05 d_05_1, d_05_2 Mar 4
6 Linear Algebra Refresher n_06 s_06 v_06 d_06_1, d_06_2 Mar 18
March 11: Midterm Exam on topics 1-5 (not 6), synchronous, your section time
7 Image Motion and Window Tracking n_07 s_07 v_07 d_07_1, d_07_2 Mar 25
8 Networks for Flow and Segmentation n_08 s_08 v_08 d_08_1, d_08_2
9 Rigid Transformations, Pinhole Camera n_09 s_09 v_09 d_09_1, d_09_2 Apr 8
10 Epipolar Geometry, 8-Point Algorithm n_10 s_10 v_10 d_10_1, d_10_2
11 Real Cameras and their Calibration n_11 s_11 v_11 d_11_1, d_11_2 Apr 22
12 Closing Thoughts s_12 d_12_1, d_12_2

COMPSCI 527, Duke University, Site based on the fluid 960 grid system