Research Group of Prof. Dr. C. Sminchisescu
Mathematical Sciences



Computer Vision (MA-INF 2201)

Summer Semester 2010

Times: Monday & Wednesday: 9 - 11am (lectures); Thursday: 8:15 - 10 am (exercises)

Place: rooms A207 (lectures), and A7a (exercises), both in Romerstr. 164

 

 

                                                                                   

Professors

 

 

      Lectures: Cristian Sminchisescu (cristian.sminchisescu at ins uni-bonn de)

 

      Exercises and projects: Fuxin Li (fuxin.li at ins uni-bonn de)

                                                   Adrian Ion (ion at ins uni-bonn de)

                            

 

     Course email: tutors.inf2201.10 at gmail.com (please, use this email for all general course inquiries)

 

 

 

Goals of the course

 

The main goal of computer vision is to make computers see, hence to infer properties of the three-dimensional world from digital images or video.  Problems in this field include identifying the three-dimensional structure of a visual scene, determining where objects in the scene are moving, and recognizing people and objects, all through analyzing information acquired with digital or video cameras.  This course provides an introduction to computer vision including topics like image representation and processing, feature detection and matching, image segmentation, motion estimation, object tracking, 3D reconstruction of rigid and articulated objects, and object recognition. Modern machine learning techniques intrinsic to reliable visual recognition, including the modeling of structured spatial and temporal dependencies, inference and prediction are also covered in the course. Along the way, different models and algorithms will be illustrated by means of practical demonstrations based on state of the art research prototypes or commercially available systems.

 

 

 

Prerequisites

 

      Foundations of Graphics, Vision and Audio (MA-INF 2101)

      Basic linear algebra (but we will review during the first exercise sessions)

      Basic vector calculus (reviewed also during the first exercise sessions)

      Fundamentals of algorithms and data structures

      Working knowledge of Matlab or C/C++

      No prior knowledge of computer vision assumed!

     

 

 

Textbooks

 

      David Forsyth & Jean Ponce: Computer Vision:  A Modern Approach, Pearson, 2002, ISBN 0130851981

      Richard Szeliski: Computer Vision: Algorithms and Applications, Springer, ISBN: 1848829345, 2011.

      Christopher Bishop: Pattern Recognition and Machine Learning, Springer, 2007, ISBN: 978-0387310732

   

      Optional: Andrew Blake and Michael Isard: Active Contours, Springer, 1998, ISBN 978-3540762171

      Optional: Richard Hartley and Andrew Zisserman: Multiple View Geometry, Cambridge, 2004, ISBN: 052154051
   

 

Administrative

 

  •  To access course materials, get username and password in class!

  •  Matlab licences can be obtained at discounted UniBonn student rates. For license information, see:
       http://www.hrz.uni-bonn.de/rechner-und-software/pc-software-lizenzen/pc-software-lizenzen-an-der-universitaet-bonn
       http://www.mathworks.de/academia/student_version/.   For online purchase (ca. 89 euro) check here.

  • For queries, and for turning in projects or assignments, please send email to:  tutors.inf2201.10 at gmail.com. All assignments (exercises and programming results) must be turned-in as hardcopies to the tutors during the exercise sessions indicated in the Syllabus. For programming assignments please send also your source code (no binaries!) by email, with subject line: "Assignment assignment_number_you_turn_in".

  • Late policy: Assignments cannot be turned in later than indicated in the Syllabus.

 

Grading

 

      50% oral exam, 50% successful exercise/assignment completion