Geometric Computer Vision

Instructor: Prof. Dr. Marc Pollefeys

Assistants: Changchang Wu and Dr. Christopher Zach
Fall 2009

Lectures: Wed 10-12 ML H37.1

Exercices: Wed 14-15 ML H37.1

 

One of the key challenges in visual computing is to measure geometric properties of objects and scenes and to recover 3D models of shape and appearance from images. Recent advances in computer vision, computer graphics and photogrammetry have resulted in effective techniques to capture models of static and dynamic scenes.   These have a wide range of applications, such as generating 3D content for web, computer games and 3D TV, but also digitizing cultural heritage or enable virtual tourism, 3D tele-presence, robot navigation and forensics.

The goal of this course is to provide students with a good understanding of how 3D object shape and appearance can be estimated from images and videos.   The main concepts and techniques will be studied in depth and practical algorithms and approaches will be discussed and explored through exercises and a course project.  

 

Course Objectives

After attending this course students should:

  1. Understand the concepts that allow recovering 3D shape from images.
  2. Have a good overview of the state of the art in geometric computer vision.
  3. Be able to critically analyze and asses current research in the area
  4. Implement components of a 3D photography system.

 

Course Topics
The course will cover the following topics a.o. camera model and calibration, single-view metrology, triangulation, epipolar and multi-view geometry, two-view and multi-view stereo, structured-light, feature tracking and matching, structure-from-motion, shape-from-silhouettes and 3D modeling and applications.

 

Target Audience
The target audience of this course are Master or PhD students, or advanced Bachelor students, that are interested in learning about geometric computer vision and related topics.  


Learning Approach

  • In the first part of the class the main geometric concepts of computer vision approaches will be covered in detail.
  • Exercises and programming assignments will be given to support this. 
  • In the second part of the class relevant papers will be discussed to explore the state-of-the-art. 
  • Students will implement components of a 3D photography system as a course project.

 

Prerequisites
The course is open to all students that have taken an introductory course in computer vision or computer graphics, or have equivalent background.  Students that are not sure if they satisfy the prerequisites should contact the instructor.

 

Course Notes
The material for this course will consist of slides, course notes and papers.  The slides will be available on-line before every class.  Part of the material is covered in course notes that are available online.  For other topics papers or notes will be made available.  Several books are also available for reference.


Slides  

  • Introduction [ppt]
  • Geometry and Camera Model [ppt]
  • Single View Metrology [ppt]
  • Feature Matching and Tracking [ppt]
  • Epipolar Geometry [ppt]
  • Shape-from-Silhouettes [ppt]
  • Stereo [ppt]
  • Structured light and active ranging [ppt]
  • Structure from motion [ppt]
  • Self-calibration and Multi-View Geometry [ppt]

Assignments  

  • see Changchang's webpage [link]

Links  


Marc Pollefeys, last updated October 28, 2009.