Probabilistic Graphical Models and their Applications Prof. Dr. Bernt Schiele

News

17.09.2021

Registration in LSF for the re-exam

Due to the fact that the winter semester is already closed, you can register for the re-exam under "Sommersemester 21 Termin 03".

The results will be transferred to the winter semester afterwards.

The date is 28.09, so you have to register / de-register before... Read more

Due to the fact that the winter semester is already closed, you can register for the re-exam under "Sommersemester 21 Termin 03".

The results will be transferred to the winter semester afterwards.

The date is 28.09, so you have to register / de-register before 21.09.

 

16.08.2021

Re-exam, 28. September

Re-exam will take place on September 28

Everybody who is interested in the re-exam should send an email to akukleva@mpi-inf.mpg.de

The exact time slots will be announced a week before the exam.

 

 

23.03.2021

PGM exam on March 25

Time slots were allocated and were sent individually to each student. 

If someone didn't get the time slot, write Anna (akukleva@mpi-inf.mpg.de).

 

06.01.2021

Projects

You are welcome to design your own project or adapt from the ones below. 

The deadline for the project is February 8. You will need to submit a report, the example you can find here.
Before January 11 choose the project and write a timeline, when you plan to do... Read more

You are welcome to design your own project or adapt from the ones below. 

The deadline for the project is February 8. You will need to submit a report, the example you can find here.
Before January 11 choose the project and write a timeline, when you plan to do what. Send this to tutors with the subject [PGM project].
Every week you are required to send a short email with what has been done and what is the plan for the next steps.

  • Predicting text from 3x4 numeric keypad inputs (T9)
    Given an input by a user (e.g. 1 2 2) build a model that can predict what the user wants to express (e.g. add or bee or bed).
    A simple solution will condition the output on the input of the last letter, yet it might be more powerful to incorporate specific user statistics as well and condition on previous (or in hindsight even following) words.
    https://en.wikipedia.org/wiki/T9_(predictive_text)

  • Image denoising with spatially dependent noise and color channels
    Multiple problems can be solved:

    • Denoising of spatially dependent noise

    • Denoising images of various types of sources (e.g. color images or x-ray medical images)

  • Image segmentation to partition an image into segments
    (for example on: archive.ics.uci.edu/ml/datasets/image+segmentation)

  • Topic models
    Assumption: data is governed by a number of latent variables (topics)
    The task would be to train a graphical topic model, for instance

    • On images (visual topics)

    • On text documents (semantic topics)

  • Probabilistic game playing agent
    Train a graphical model representing a stochastic game, such as Blackjack (one or more players)

  • Local statistical shape model:
    Build local statistical shape models for each part of the body and define a graph that connects the parts to produce coherent shapes. The graph structure can be a tree or with loops.
    The end goal is then to sample from the graph to produce realistic shapes with more variation than using a global shape model.
    http://smpl.is.tue.mpg.de/ (server might be down for 1-2 days)
    https://ps.is.tuebingen.mpg.de/publications/smpl-2015

  • Shape registration using distributed inference: 
    Given two 3D shapes (for example of humans), split the shapes into parts and find the alignment between them. In the lecture exercises you have found the registration by treating the shape as a whole. In this project the idea is to consider the shape as a union of weakly connected parts. The goal is to find the optimal alignment by designing an energy function that can be decomposed into unaries (part registration cost) and pairwise terms (part to part cost).

  • 3D pose and shape estimation using particle message passing: 
    Using a variation of particle filters for graphs track the human pose and shape of people in images. The human pose of the person is represented in 3D as a collection of loosely connected parts.

  • Super-resolution.

    Given low-res and high-res training images your task is create an algorithm that output high-res image for the input low-res images.

    For the reference you can use this paper: 

    http://people.csail.mit.edu/billf/project%20pages/sresCode/Markov%20Random%20Fields%20for%20Super-Resolution.html

  • Background subtraction.

    Your task is to create algorithm on images  or videos to separate background from foreground. 

    As dataset for images you can use  CIFAR or Mini-ImageNet.

    For videos Weizman dataset. ( videos you can dowload from here: http://www.wisdom.weizmann.ac.il/~vision/SpaceTimeActions.html)

    As a refernece for the algorithm you can use GrubCat alg. https://cvg.ethz.ch/teaching/cvl/2012/grabcut-siggraph04.pdf

  • Semantic segmentation.

    The task is to segment images into semanticaly meaningful segments. 

    For the task use Pacal VOC 2010 dataset.

    Reference paper: http://jamie.shotton.org/work/publications/ijcv07a.pdf

Sample code to work with 3D human shapes:

03.01.2021

Assignment 3: deadline extension

Due to requests, the deadline for assignment 3 has been extended by a day to 5.01.21 (Tuesday) 11:59pm and an additional office hour (by Apratim) scheduled for 10am - 12pm on 4.01.21 (Monday).

22.12.2020

Exam

Examination will be oral.

Please register with LSF so the exam counts.

We suggest two possible dates:  11 February and 25 March. The preferred date should be send as soon as possible by email to one of the tutors, and no later than 4 days before the... Read more

Examination will be oral.

Please register with LSF so the exam counts.

We suggest two possible dates:  11 February and 25 March. The preferred date should be send as soon as possible by email to one of the tutors, and no later than 4 days before the exam.

Final time slots will be assigned couple of days prior to the exam and send individually to students. 

 

14.12.2020

Assignment 3

The third assignment has been released (in Materials). Due date: 4th Jan (2021), 11:59 pm

09.12.2020

Assignment 2: deadline extension

The deadline for the second assignment is extended till 13th December 11:59. 

27.11.2020

Assignment 2

The second assignment has been released (in Materials). Due date: 10th Dec, 11:59 pm
13.11.2020

Assignment 1

The first assignment has been released (in Materials). Due date: 25th Nov, 11:59pm.

11.11.2020

Office Hours for Tutors

Office hours will be virtual via Zoom. 

Apratim Bhattacharyya (Wednesday 11am - 1pm)

Anna Kukleva (Tuesday 2 pm - 4 pm) 

 

zoom links in the materials

10.11.2020

[Pgm_annoncements] mailing list

We have added all course participants to the mailing list: https://lists.mpi-inf.mpg.de/listinfo/pgm_announcements

If due to any reason, you have not received the welcome email, you can subscribe by clicking on the link above.

This list will be used for... Read more

We have added all course participants to the mailing list: https://lists.mpi-inf.mpg.de/listinfo/pgm_announcements

If due to any reason, you have not received the welcome email, you can subscribe by clicking on the link above.

This list will be used for important announcements e.g. zoom links for lectures.

Please send an email to the tutors if you have any questions.

03.11.2020

Zoom link for the first lecture

 

https://zoom.us/j/94066494694?pwd=dDdLUElrVCswSDdDOWx4UmR4RVJ1UT09

Meeting ID: 940 6649 4694
Passcode: 547442

03.11.2020

Zoom link for the first exercise

 

https://zoom.us/j/97145007820?pwd=Yk9ZK1ZzTUdKWmljOWprK0Y2M0MxZz09

Meeting ID: 971 4500 7820
Passcode: 535352

Matlab installation: https://www.hiz-saarland.de/dienste/software-lizenzen/mathworks/ (with UdS campus license)
Instruction: Tutorial

Show all
 

Probabilistic Graphical Models and their Applications

Overview

This course will introduce the basic concepts of probabilistic graphical models. Graphical Models are a unified framework that allow to express complex probability distributions in a compact way. Many machine learning applications are tackled by the use of these models, in this course we will highlight the possibilities with computer vision applications.

 

The main goal of the class is to understand the concepts behind graphical models and to give hands-on knowledge such that one is able to design models for computer vision applications but also in other domains. Therefore the lecture is roughly divided in two parts: learning about graphical models and seeing them in action.

 

In the first part of the lecture we will discuss the basics of solving these models, eg. for special kinds of graphs where efficient exact inference is possible and approimate methods for the general case. In the second part we will then discuss prominent applications for both low- and high-level computer vision problems. Some examples are statistical models of images (eg denoising), body pose estimation, person tracking, object detection and semantic image segmentation.

 

The exercises will be a mix of theoretical and practical assignments.



Privacy Policy | Legal Notice
If you encounter technical problems, please contact the administrators