Arrangements for the exercise sessions
This course contains 2 hours of "compulsory" exercises per week for 7 weeks, for a total 14 hours. The exercises are "compulsory" in the sense that the points you collect from tasks you are prepared to present at the exercises directly correspond to 50% of the course grading. Attending each session is not necessary, but highly recommended.
Each exercise session contains several tasks. A student will be awarded 1 - 3 points for each task depending on the difficulty of the task. The number of tasks per week varies. On average each session will provided 4 (+-1) points. At the end of each exercise session the tasks for the next week will be shortly presented and students will have possibility to ask questions about the tasks.
The tasks for the next week session will be published by the current week session in order to give students the opportunity to go through the tasks and prepare the questions.
The tasks must be solved before the exercise session to get the points. In the beginning of each exercise session, you must mark the completed tasks in a list provided by the teaching assistant. After that the solution for each task will be demonstrated by a student who has marked
that particular task. The criteria for picking are influenced by the participation of the students to the session's discussions (don't participate actively and you will be asked to participate).
If a student has marked a task but cannot explain the solution (e.g. as a result of copying), they will be penalized by -5 points.
You can collect points only for those tasks that you are ready and prepared to present in class. Don't be late from the class or you will miss the possibility of marking the first (or consequent) task!
If you cannot attend an exercise session but wish to complete some tasks, you can return the answers by e-mail. Answer is generally runnable Matlab code + answers to all the questions of the task. However, this is only allowed for very good reasons and must always be agreed with the TA during the previous week before the exercises. Then the solutions must be sent well before the corresponding exercise session (in practise, latest on Monday morning 8:00 before the Monday afternoon exercise. This is an hard deadline). Skipping all sessions by sending the answers by email is not possible. Special cases must be discussed with TA and professor.
No points are given for the first week exercises. Matlab is used for numerical computations during the course. For those not familiar with Matlab, the following pages may be found helpful:
- Getting started with Matlab (from MathWorks)
- Matlab primer
- Matlab tutorial
- A practical introduction to Matlab
02.01. Exercise 0: Matlab primer, arrangements (no points from this session)
09.01. Exercise 1: Image filtering
16.01 OR 18.01. Exercise 2: Feature extraction and matching
23.01 OR 25.01 Exercise 3: Camera models and calibration
30.01 OR 01.02. Exercise 4: Localization and Tracking
06.02 OR 08.02. Exercise 5: Stereo Vision
13.02 OR 15.02. Exercise 6: 3-D reconstruction