CPSC 4820/6820 Applied Computer Vision

Description: Students are introduced to computer vision techniques starting with basic digital image processing. Concepts covered will include image convolution, edge detection, and the Fourier and Wavelet transforms. The course will also introduce students to video processing, including camera calibration and concepts related to Augmented Reality, including marker detection and camera homography.
Prerequisites: CPSC 2120.
Required texts: Szeliski, Richard, Computer Vision: Algorithms and Applications, (c) Richard Szeliski, 2010.
Recommended texts: Glassner, Andrew, Principles of Digital Image Synthesis, (c) Andrew Glassner, 2014.
Supplemental notes: None.
Professor: Dr. Andrew Duchowski
Office: McAdams 309, 656-7677, duchowski@clemson.edu
Office hours: By appointment.
Objectives: To gain a basic understanding of computer vision algorithms and then to (install and) use OpenCV.
Laboratory content: This is not a lab course.
Programming Assignments35%
Term paper10%
Final Presentation5%
Midterm Exam20%
Final Project30%
Assignments: Unless otherwise specified, these are individual assignments, and must be strictly your own work and are not to be shown to anyone else.
Attendance: Roll will be taken for the first one or two weeks while the class roll fluctuates. However, attendance is not required. Absence, excused or not, does not change the responsibility for assigned work. Tests missed due to excused absences will normally result in the test not being counted in the average grade (i.e., there will normally be no makeup tests). An unexcused absence from a test will normally result in a grade of zero for that test. Students are expected to give at least one week advance notice for excused absences.
Academic dishonesty: The University policies on academic dishonesty apply. Publicly-available code or other material may be freely used if appropriately attributed. Each student is responsible for protecting his or her files from access by others. Work that is essentially the same and submitted without proper attribution is considered to be a violation of academic dishonesty policy by all those submitting the work, regardless of who actually did the work.
Class cancelation: Students are expected to wait for 15 minutes after the class beginning time before leaving if the instructor is late.
Assignment late policy: Late assignments will be accepted but points will be deducted according to the formula (3n)3 where n is the number of days late. Example: assuming assignments are due on Wednesday, the point deduction is as follows:
Max points possible Day received Days late
100Wed.0 (due date)
73Thu.1 day late
0Fri.2 days late

Late assignments will receive lowest priority for grading and returning.

Topical outline:
  • Image I/O
  • Image Convolution with Sobel (FIR) Filter
  • Edge Detection
  • Fourer Transform
  • Wavelet Transform
  • Video I/O
  • Camera Calibration / Homography
  • Fiducial Marker Detection & Tracking
  • Face Detection & Tracking
  • Hotelling (Karhunen-Loève) Transform
  • Principal Component Analysis (PCA)
  • Computational Photography
  • High Dynamic Range (HDR) Tone Mapping with Bilateral (IIR) Filter