Computer Vision 2

Carsten Rother, Michael Yang, Alexander Kirillov, Summer Semester 2016

 

The lecture computer vision 2 consists of lecture and practical exercises. The lecture is partitioned in three parts. The first part looks at general recognition tasks and scene understanding. In the second part, deep learning methods, e.g. CNN, are introduced. In the final part, Markov random fields are described in details and algorithms for segmenting objects in images are considered.

 

Lectures: Friday, 13:00 Р14:30, APB E006, Begin: 08. April 2016.
Exercise: Thursday, 11:10 – 12:40, APB E010 APB E069, Begin: 14. 21. April 2016.
Prerequisites: “Computer Vision 1”, good knowledge of maths (linear algebra, optimization), programming (C++).
Credits: 2/2/0, oral exam, Enrollment: jExam, Attendees: max. 60.
Note: Lectures are held in English with slides in English. There is one course book: “Computer Vision: Algorithms and Applications” by Richard Szeliski which can also be found online: http://szeliski.org/Book.

 

Note on the oral exam:
The exercises are part of the exam; if you obtain at least half the achievable number of points in the exercises the questions that regard the exercise will concentrate around what you did, otherwise they will cover the whole set of exercise tasks.
Scripts:
Lectures: (slides available around time of lecture)

08.04 Introduction+Detection (Carsten, Michael) slides
15.04 Scene Understanding (Michael) slides
22.04 Image Categorization (Michael) slides
29.04 Deep Learning 1 (Alexander) slides
06.05 no lecture
13.05 Deep Learning 2 (Alexander) slides
27.05 Deep Learning 3 (Alexander) slides
03.06 Deep Learning 4 (Alexander) slides
10.06 Graphical Models 1 (Carsten) slides
17.06 Graphical Models 2 (Carsten) slides
24.06 Segmentation 1 (Carsten) slides
01.07 Segmentation 2 (Carsten) slides
08.07 Other Segmentation Approaches (Dmitrij) slides
15.07 Graphical models: statistical inference and learning (Dmitrij) slides

 

Exercise:
There will be 3 topics for exercises. Each will have different tasks.

Exercise 1: Face detection using a SVM and HOG-Descriptor task , data , optional data. The exercise is in room APB E069. You have to show your results until the 19th 26th of May.
Exercise 2: CNN task and data You have to show your results until the 16th 23th of June.
Exercise 3: Image segmentation task