Persönlicher Status und Werkzeuge

Home Teaching Winter Semester 2012/13 Statistical Methods and Optimization in Computer Vision

Statistical Methods and Optimization in Computer Vision

WS 2012/13, TU München

Lecture

Location: Room 02.09.023
Date: Thursday, starting at October 25th
Time: 11.15am
Lecturer: Dr. Claudia Nieuwenhuis
ECTS: 4
SWS: 3

Seminar

Location: Room 02.09.023
Date: 6.11.2012, every other week
Time: 10.15am
Lecturer: Eno Töppe

The course will be held in English.

There will be no lecture on January 31st!

Contents

Statistics is the foundation of many powerful tools in computer vision. This lecture will cover a number of widely used and important techniques for the analysis of images containing large amounts of data. We will discuss a selected number of approaches concerning their mathematical theory and implementation details. Topics will cover

- necessary basics in statistics, e.g. distributions, conditional distributions, marginal distributions, cumulative distribution functions, MAP, MLP, Bayes theorem, hypothesis tests

- subspace methods such as principal component analysis, idependent component analysis, linear discriminant analysis, e.g. with application to face recognition

- density estimation and sampling methods such as Parzen density estimation and particle filtering, e.g. with application to image segmentation and tracking

- learning and classification approaches such as Support Vector Machines, Neural Networks, Graphical Models and Dictionary Learning

- optimization methods such as variational approaches, PDEs, MRFs

Slides and Exercises

The lecture slides and exercise sheets can be downloaded here.

Last edited 23.01.2013 14:15 by nieuwenh