The legacy resources below are left in place for your reference only. For W25 students, please access Assignments in the sidebar for your lab assignments.

Lab assignments

1. Motion Estimation and Tracking
   1. Optical flow
   2. 2D Video tracking in Matlab
   3. (May be optional) Video tracking in MTF
2. Stereo and Geometry I
   1. 2D projective geometry and homography estimation
   2. 3D projective geometry and stereo
3. Geometry II and SFM
   1. (Likely optional) Stereo geometry and augmented reality
   2. Structure from motion
4. (May be optional) Reflectance and photometric stereo

The lab component will focus on reinforcing a principled understanding of the geometry and mathematics of vision rather than real-time implementation. Therefore, labs will be done in the high level scripted mathematics language Matlab. Critical real-time components such as the video pipeline, capture and tracking subroutines will be provided through our MTF system, and accessed in Matlab using dynamically linked subroutines. I have used this 2-level approach with Matlab linked to compiled code in the University of Rochester computer vision course and found it to work well.

Lab Tutorials

1. mexMTF: the image capture software
2. Intro. to Matlab, capturing images

Obsolete documentation on mexVision, the previously used image capture pipeline

To get access to the Vison lab (CSC 1-05), please see Sharon Bell at the Athabasca office front desk. Tell her you are enrolled in CMPUT 428/615 and need access to the lab.

Remote access to the lab machines: ulxx.cs.ualberta.ca, where xx is 01-10