Instructor: Jane Mulligan

CSCI 5722: Computer Vision is an introductory graduate level course on exploiting digital images as a sensor, to extract useful information about the world. (We will NOT cover image processing techniques such as MPEG encoding)

Prerequisites: Linear algebra, basic trigonometry, and probability. Knowledge of MATLAB is a benefit.

syllabus.pdf

Text: None required; technical readings will be assigned and made available on this site.
Useful resource texts:

• Computer Vision -- A modern approach -- David Forsyth and Jean Ponce (also see this pointer).
• Image Processing, Analysis and Machine Vision -- Milan Sonka, Vaclav Hlavac and Roger Boyle.
• Computer Vision -- Dana Ballard and Christopher Brown a classic text which has been digitized and made available online.
• Computational Vision -- Hanspeter A. Mallot John S. Allen another online text from MIT press (2000).

Topics :

• Image Formation:
  • Cameras and projection
  • Light and surfaces
  • Representations
• Calibration:
  • Strong
  • Weak
• 2D Vision
  • Filters
  • Binary Images
  • Features
  • Edge Detection
  • Texture
  • Shape
• Segmentation
  • Clustering
  • Model Fitting
  • Probabilistic
• 3D Vision
  • Stereo
  • Shape from X
  • 3D Data
• Dynamic Sequences
  • Optical Flow
  • Tracking
  • Structure from Motion
• Recognition
  • Representations
  • Appearance Based
• What's New

Grading:

• Class participation: 10%
• Assignments: 50%
• Project: 40%

Assignments

1. Photometric Stereo
2. Multiscale Feature Extraction and Matching
3. Structure and Motion, pdf.
4. Background subtraction, pdf.

Project , Presentation Schedule .

Final grades.

Notes etc

• Lectures will be posted below after each class.
• Introduction (Reading Chapter 1 Computational Vision --Mallot and Allen ).
• Cameras and Projection
• Color I
• Radiometry , Physics-based vision example.
• Color II .
• Wei Xu's guest lecture
• Linear Filters .
• Edges, Assignment slides (see assignment page link above).
• Features .
• Texture and Pyramids .
• Tracking.
• Segmentation.
• Segmentation and Fitting
• Camera Calibration, Matlab Calibration Toolbox, Zhang's ICCV'99 Paper.
• MultiView Geometry.
• Stereo.
• Structure from Motion.
• Recognition by Template Matching.
• Relations between Multiple Templates.
• Image Blending.
• Image-based retrieval, City-Scale Location Recognition, Schindler et al., Content-based Multimedia Information Retrieval: State of the Art and Challenges, Lew et al.