Direkt zum Inhalt springen
Computer Vision Group
TUM Department of Informatics
Technical University of Munich

Technical University of Munich

Menu
Home Teaching Summer Semester 2019 Probabilistic Graphical Models in Computer Vision (IN2329) (2h + 2h, 5 ECTS)

Probabilistic Graphical Models in Computer Vision (IN2329) (2h + 2h, 5 ECTS)

Announcement:
There was an error in the exercise sheet 9, which is corrected now (2019.07.09 16:00).
The last lecture on 22.07 will be on deep Boltzmann machines presented by Yuesong Shen.
There will be NO tutorial on Wednesday, 12.06.2019. Sheet5 should be submitted on 17.06.
There will be NO lecture on Wednesday, 24.04.2019.


Introduction:
Several problems in computer vision can be cast as a labeling problem. Typically, such problems arise from Markov Random Field (undirected graphical models), which provide an elegant framework of formulating various types of labeling problems in vision. Under certain assumptions some "nice" MRF models can be solved in polynomial time, whereas other approaches are NP hard. We will see both efficient algorithms for solving the "nice" problems and relaxation strategies for the "hard" ones.



We will cover the following topics.

- Graphical model representations:

  • Bayesian network
  • Markov network
  • Conditional random field
  • Factor graph
  • Exponential family

- Graphical model inference:

  • Variable elimination
  • Junction-tree algorithm
  • Belief propagation
  • Graph cut and move-making extensions
  • Linear programming relaxation

- Approximative inference techniques:

  • Loopy belief propagation
  • Mean field, principle of variational inference
  • Sampling methods

- Graphical model learning:

  • Maximum-likelihood estimation
  • Expectation-maximization algorithm
  • Structured SVM

- Further topics if time allows

We will implement some of the discussed methods in Python.


Lecture

Location: Room 02.09.014
Time and Date: Monday 16:15 - 18:00
Start: April 29th, 2019
Lecturer: Dr. Tao Wu
The lecture is held in English.

Exercise

Location: Room 02.09.023
Time and Date: Wednesday 12:15 - 14:00
Start: May 08th, 2019
Organization: Yuesong Shen, Zhenzhang Ye

The exercise sheets consist of two parts, theoretical and programming exercises.

Exercise sheets will be posted every Monday and are due a week later. You will have one week to do the exercises.
Please submit the programming solutions as a zip file with filename "matriculationnumber_firstname_lastname.zip" only! containing your code-files (no material files) via email to pgm-ss19@vision.in.tum.de, and hand in the solutions to the theoretical part in Monday's lecture. We will give you back the corrected sheets on Wednesday when we discuss them in class.
Please remember to write clean, commented(!) code! You are allowed to work on the exercise sheets in groups of two students.
The first exercise sheet that counts to the exam bonus is exercise sheet 1, which is due on 13th May.

The exercise sheets can be accessed here.

Exam Bonus

To achieve the bonus, you have to meet two requirements:
1. get at least 75% grades of all exercises totally, i.e. sum of the grades you get in all exercises divided by the grades of all exercises (without bonus) should be >= 0.75
2. present your theoretical solution during tutorial at least once in this semester.
You cannot improve either 1.0 or >4.0.

Exam

Date: August 05th, 08:30 - 09:45.
Place: 102, Interims Hörsaal 2 (5620.01.102)
The final exam will be written. No cheat sheet is allowed.

Lecture Materials

Course material (slides and exercise sheets) can be accessed here.

Send an email to pgm-ss19@vision.in.tum.de if you need the password.

Rechte Seite

Informatik IX
Chair of Computer Vision & Artificial Intelligence

Boltzmannstrasse 3
85748 Garching

info@vision.in.tum.de