Instructor: Alper Yilmaz, PhD
Office: Bolz Hall 214C
Lecture Day&Time: Tuesday 1:30pm-4:18pm

Cource Book:
R. Hartley and A. Zisserman: Multiple View geometry in Computer Vision, Cambridge University Press, 2000.
Google book search link to text book

Cource Objective: GS830 provides students with both theoretical and practical aspects of implicit geometric relations between multiple images of a scene taken from different viewpoints. Several examples will be provided throughout the course to provide the students the tools to apply these cocepts on real world images.

Lecture Slides:

• Lecture 1 (3/27/2007)
  • Introduction and Course Layout
• Lecture 2 (3/29/2007)
  • Projective geometry
  • Transformations
• Lecture 3 (4/3/2007)
  • Projective geometry
  • Recovery of affine and metric properties
  • Homework assignment
• Lecture 4 (4/5/2007)
  • Projective geometry in 2D
  • Homework assignment
• Lecture 5 (4/10/2007)
  • Projective geometry in 3D
• Lecture 6 (4/12/2007)
  • Estimation of 2D Projective Transformations
  • Programming assignment
    • Input images for assignment 1: (Set 1) (Set 2)
    • Input images for mosaic: (Set 1) (Set 2)
• Lecture 7 (4/24/2007)
  • Estimation of 2D Projective Transformations
  • Programming assignment
    • Input images for assignment
• Lecture 8 (4/25/2007)
  • Camera models
• Lecture 9 (4/26/2007)
  • Camera models
  • Estimation of projection matrix
• Lecture 10 (4/30/2007)
  • Estimation of projection matrix
  • Single view geometry
• Lecture 11 (5/3/2007)
  • Single view geometry
• Lecture 12 (5/4/2007)
  • Single view geometry
  • Multiple view geometry
• Lecture 13 (5/8/2007)
  • Multiple view geometry
  • Direct reconstruction
• Lecture 14 (5/10/2007)
  • 3D reconstruction
  • Computation of F
• Lecture 15 (5/15/2007)
  • Computation of F
• Lecture 16 (5/17/2007)
  • Structure computation
  • Scene planes and homographies
• Lecture 17 (5/22/2007)
  • Trifocal tensor
• Lecture 18 (5/24/2007)
  • Factorization
  • Project discussions
• Lecture 19 (5/29/2007)
  • Projects
• Lecture 20 (5/31/2007)
  • Projects

Due date: (4/17/2007). See Lecture Notes 4

Due date: (4/26/2007). See Lecture Notes 6

Due date: (5/08/2007). See Lecture Notes 7

Due date: (4/10/2007). See Lecture Notes 3

Due date: (4/12/2007). See Lecture Notes 4

Due Date & Time: 6/8/2007 3pm

Deliverables:

• Weekly progress reports. Including half a page to one page of text with 11pt Times New Roman single line.
• Paper in IEEE two column format. 6 pages of text. Times New Roman 11pt, single space.
• Half a page peer-evaluation
• Source code on a CD with written names of the contributors
• 25 minute presentation
• Program demo in the class

Object recognition using view-geometric constraints
Gabor Barsai, Yeon Yeun, Sudhagar Nagarajan

3D Object reconstruction using level sets
Panu Srestasathieren, Ekta Amar, Sagar Desphante

Object grouping using view-geometric relations
Po-Lun Lai, Shaojun He, I-chieh Lee

P. Kovesi: Matlab toolbox for various multi-view geometry based algorithms

A. Zisserman: Matlab functions for multi-view geometry

A. Heyden: Tutorial on Multiple View Geometry

A. Mundy: Tutorial on Projective Geometry for Machine Vision

S. Heuel: Projective Geometry for Grouping and Orientation Tasks

W. Foerstner: Uncertainty, Testing and Estimating of Geometric Parameters

Z. Zhang: Parameter Estimation Techniques: A Tutorial with Application to Conic Fitting