Philippos Mordohai
Assistant Professor
Department of Computer Science
Stevens Institute of Technology

Office: Lieb 215
Phone Number: +1 201 216 5611
E-mail: mordohai_at_cs.stevens.edu

CS 532: 3D Computer Vision

Fall 2015



Homepage

Location
TBD

Time
Wednesday 6:15-8:45 PM.

Office Hours
Tuesday 5-6 and by appointment.

Pre-requisites
Programming, data structures matrix operations. Instructor’s permission.

Syllabus

Textbooks
Parts of the following resources, both of which are available free of charge online, will be used. The following book is an invaluable resource for all researchers in topics covered by this course, but it is optional.
  • Richard Hartley and Andrew Zisserman, Multiple View Geometry in Computer Vision, Second Edition, Cambridge University Press, 2004.
Evaluation
  • Seven homework assignments (70%). Homeworks will be assigned every other week mainly focused on implementing concepts discussed in class.
  • Weekly quizzes (15%).
  • Final exam (15%).

Resources

The following links should be useful in case you need to refresh your math or Matlab knowledge. Class Schedule

Week 1: Image formation, homogeneous coordinates (Szeliski Ch. 2, Hartley and Zisserman slides)
Lecture 1 slides (pdf)

Week 2: Homography estimation, RANSAC, two-view geometry (Szeliski Ch. 11, Hartley and Zisserman slides)
Lecture 2 slides (pdf)
Homework 1 (pdf and ppm image) is due on Sep. 16.

Week 3: Fundamental matrix estimation, Binocular stereo, matching criteria (Szeliski Ch. 11)
Lecture 3 slides (pdf)

Week 4: Stereo Matching confidence, Feature extraction (Szeliski Ch. 7, Hartley and Zisserman slides)
Lecture 4 slides (pdf)
Homework 2 (pdf and data) is due on Oct. 7.

Week 5: KLT tracking
Lecture 5 slides (pdf)

Week 6: Simultaneous Localization and Mapping, Kalman filtering (Green notes, Welch and Bishop tutorial)
Lecture 6 slides (pdf)
Homework 3 (pdf) is due on Oct. 21.

Week 7: Structure-from-Motion (Szeliski Ch. 7, notes from several sources)
Lecture 7 slides (pdf)

Week 8: Photo-tourism and multi-view stereo (part I) (notes)
Lecture 8 slides (pdf)
Cloth3 images and ground truth
Homework 4 (pdf) is due on Nov. 4.

Week 9: Structure-from-Motion (part II) (Szeliski Ch. 7, notes from several sources) (Friday, October 30)
See Lecture 7 slides above.

Week 10: Multi-view stereo (part II) and silhouette-based modeling; introduction to computational geometry and convex hull estimation in 2D and 3D (notes; Mount Lec. 21 and 3)
Lecture 9-10 slides (pdf)
Homework 5 data
Homework 5 (pdf) is due on Nov. 18.

Week 11: Line intersection, introduction to polygon triangulation (Mount Lec. 3, 5 and 6)
Lecture 11 slides (pdf)

Week 12: 3D mesh representation (Mount Lec. 22)
Lecture 12 slides (pdf)
Gargoyle ply model
Homework 6 (pdf) is due on Dec. 2.

Week 13: Unorganized point clouds, normal estimation, invariant descriptors for 3D data (Mount Lec. 16 and notes)
Lecture 13 slides (pdf)
Homework 7 data
Homework 7 (pdf) is due on Dec. 11.

Week 14: Delaunay triangulations and Voronoi diagrams (Mount Lec. 11, 12 and 13)
Lecture 14 slides (pdf)


Resources