A Computer Vision Sampler

This course explores concepts in visual recognition, visual motion analysis, and visual reconstruction through a sample of techniques. There will be lectures, homework, and two exams.

The class meets on Tuesdays and Thursdays from 8:30 am to 9:45 am in room B101 (the Love Auditorium) of the Levine Science Research Center (LSRC) Building.



  • The midterm exam is on Thursday, March 9, from 8:30 am to 9:45 am.
  • The final exam is on Friday, May 5, from 2 pm to 4 pm (two hours, not three).
Both exams will be in the lecture room for this course.


We will use Ed STEM to communicate with each other outside of class or office hours. Please access Ed STEM through Sakai, not directly, for appropriate access privileges.

[Image by B. Palac from Wikimedia used under the Creative Commons Attribution-Share Alike 4.0 International license.]

A typical computer vision task is to segment images, that is, to specify which of a predetermined set of categories each image pixel belongs to. Categories are as follows in this example: Yellow: building. Red: vehicle. Green: vegetation. Purple: road. Blue: none of the above.

Other tasks may require knowing where certain objects are in a 3D reference system tied to the camera, or which way and how fast something is moving in a video.

There are plenty other applications and tasks in computer vision. We look at some of these in the first lecture.

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