Lecture with Exercises: Computer Vision

Lecturer: Prof. Andres Bruhn
Coordinator of Tutorials: Daniel Maurer, Michael Stoll

Winter Term 2018 / 2019
Lecture with Exercises (3+1SWS)
Language: English


Lecture: Wednesday, 11:30 - 13:00, Computer Science Building, Lecture Hall V38.03.
Lecture: Friday, 9:45 - 11:15, Computer Science Building, Lecture Hall V38.03 (bi-weekly).
Exercises: Friday, 9:45 - 11:15, Computer Science Building, Lecture Hall V38.03 (bi-weekly).

Description

This class gives an in-depth introduction into the field of computer vision. It consists of four parts. In the first part, characteristic image features and feature descriptors are discussed that are typically used for recognition or matching, The second part of the course is dedicated to motion and stereo. Fundamental goals in this context are the recovery of the 3-D structure and the 2-D motion of objects from a recorded scene. The third part of the class is dedicated to segmentation - the subdivision of an image into different semantically meaningful regions. Here, the focus lies on contour based methods that are among the leading techniques in the field. Finally, the fourth part deals with pattern recognition the classification of features into different groups. In this context, basic approaches for the reducing the dimensionality of features and strategies for supervised learning are discussed.

This class requires undergraduate knowledge in mathematics (e.g. ''Mathematik für Informatiker und Softwaretechniker", "Numerische und stochastische Grundlagen der Informatik"). The previous attendance of the class "Imaging Science" is recommended. Lectures will be in English.

This class is particularly useful for those students who wish to pursue a bachelor or master thesis in our group.

Course Material

Part 1: Features and Descriptors
 
Lecture 01 17.10.2018 Introduction
Lecture 02 19.10.2018 Features and Descriptors I: Linear Diffusion,  Scale Space
Lecture 03 24.10.2018 Features and Descriptors II: Image Pyramids,  Edges and Corners
Lecture 04 26.10.2018 Features and Descriptors III: Hough Transform, Invariants
Lecture 05 31.10.2018 Features and Descriptors IV: Texture Analysis
Exercise 01 02.11.2018  
Lecture 06 07.11.2018 Features and Descriptors V: Scale Invariant  Feature Transform


Part 2: Motion and Stereo

Lecture 07 09.11.2018 Image Sequence Analysis I: Local Methods
Lecture 08 14.11.2018 Image Sequence Analysis II: Motion Models, Tracking
Exercise 02 16.11.2018  
Lecture 09 21.11.2018 Image Sequence Analysis III: Variational Methods
Lecture 10 25.11.2018 3-D Reconstruction I: Camera Geometry
Lecture 11 28.11.2018 3-D Reconstruction II: Epipolar Geometry, Stereo Matching
Exercise 03 30.11.2018  
Lecture 12 05.12.2018 3-D Reconstruction III: Shape-from-Shading
 

Part 3: Segmentation

Lecture 13 07.12.2018 Foundations I: Isotropic Nonlinear Diffusion
Lecture 14 12.12.2018 Foundations II: Anisotropic Nonlinear Diffusion
Lecture 15 14.12.2018 Segmentation I: The Mumford/Shah Mode
Exercise 04 19.12.2018  
Lecture 16 21.12.2018 Segmentation II: Continuous Scaled Morphology,  Shock Filters
No Lecture 26.12.2018 Christmas Holidays
No Lecture 28.12.2018 Christmas Holidays
No Lecture 30.12.2018 Christmas Holidays
No Lecture 02.01.2019 Christmas Holidays
No Lecture 04.01.2019 Christmas Holidays
Lecture 17 09.01.2019 Segmentation III: Mean Curvature Motion
Lecture 18 11.01.2019 Segmentation IV: Self-Snakes, Active Contours
Exercise 05 16.01.2019  
 

Part 4: Pattern recognition

Lecture 19 18.01.2019 Pattern Recognition I: Basics, Terminology
Lecture 20 23.01.2019 Pattern Recognition II: Bayes Decision Theory
No Lecture 25.01.2019 cancelled
Lecture 21 30.01.2019 Pattern Recognition III: Parametric Techniques, Density Estimation
Exercise 06 01.02.2019  
Lecture 22 06.02.2019 Pattern Recognition IV: Non-Parametric Techniques
Lecture 23 08.02.2019 Pattern Recognition V: Dimensionality Reduction, Outlook


Assignments
Math Sheet issued 17.10.2018 no grading
Programming Sheet issued 17.10.2018 no grading
Exercise 01 issued 26.10.2018 submission 02.11.2018
Exercise 02 issued 09.11.2018 submission 16.11.2018
Exercise 03 issued 23.11.2018 submission 30.11.2018
Exercise 04 issued 12.12.2018 submission 19.12.2018
Exercise 05 issued 09.01.2019 submission 16.01.2019
Exercise 06 issued 25.01.2019 submission 01.02.2019
 

Links

Link to "Computer Vision" in the Campus system (Lecture)
Link to "Computer Vision" in the Campus system (Exercise)
Link to "Computer Vision" in ILIAS

Remarks
  • The programming exercises are designed for Linux. We only guarantee that the code works under Linux.
  • If possible, we will provide executables for Windows. You may use them at your own risk.
  • Sometimes we may have to provide object files for copyright reasons. Object files will only be provided for Linux.
  • If there are any problems reading in and writing out files under Windows, adding the binary flag may help, i.e. inimage = fopen(in,"rb");