# Computer Vision Exercises

Summary

Two exercise sessions:

Monday: 12:15 - 13:00, Room 00.151-133 (NEW! MORE SPACE!)

Wednesday: 10:15 - 11:00, Room: 09.150

# Description

In the exercises, we will have theoretical and practical assignments. The theoretical assignments aim at a deeper understanding of the principles tought in the lecture. The practical part offers hands-on examples to the participants.

Exercises are not mandatory, but students are strongly encouraged to work through the exercise sheets. Usually we will have one exercise sheet per week. Solutions will be discussed in the exercise at the stated due date.

In case you have questions or need feedback on your work, do not hesitate to contact Simone!

All mathematical concepts that are used in the lecture and exercises should have been covered in the basic lectures on mathematics. Even if you don't remember, it should be possible for you to catch up on-the-fly.

However, if you are interested in refreshing your mind, these are the main concepts used (Disclaimer: probably not an exhaustive list, just a general guideline):

- Linear Algebra, Matrix-/Vector operations
- (Gaussian) Distributions
- Convolution
- conditional probability, Bayes Law

- Projections (Orthographic, perspective, ...)
- homogeneous coordinates
- parameterized lines/surfaces
- coordinate transforms

## NEWS

23.04.2014:

Beginning next week, the exercise slides and solutions will be password protected. The login data will be provided in the lecture and exercises.

02.07.2014:

Change of rooms: The exercise rooms in the computer science building will be closed next week, so the last exercise sessions will be held at room 01.255-128 (new computer science building)!!

# Slides

Here you will find the current slides:

# Worksheets

Here you will find the current worksheets and data:

### Exercise sheets:

OVERVIEW of the main points covered in the exercises

# Coding Setup

We use C++ and the OpenCV framework in the assignments.

Have a look at the current OpenCV documentation.

### CIP-Pool

The easiest way is working at the Computer Science CIP-Pools. OpenCV is ready to use there.

First, set the following environmental variables:

export PKG_CONFIG_PATH=/local/opencv/lib/pkgconfig/
export LD_LIBRARY_PATH=/local/opencv/lib

Then, the compiler flags needed to build with OpenCV should be available through this command:

pkg-config opencv --cflags --libs

You can use the output on the command line on the fly:

g++ `pkg-config opencv --cflags --libs` -o example example.cpp

For more convenience, write a Makefile.

### Windows

Use, e.g., the free Express Editions of Visual Studio.

### Linux

Our Linux-Computer-Vision-Guys like to use QT Creator, as it works well with CMake.

### Both

Get OpenCV, use CMake to create your projects (compare Slides1, see above Slides section).