IT-Universitetet i København
  Tilbage Kursusoversigt
Kursusnavn (dansk):Digital filmbehandling 
Kursusnavn (engelsk):Digital Post Processing of Cinema Movies 
Semester:Efterår 2003 
Udbydes, multimedieteknologi (mmt) 
Omfang i ECTS:7,50 
Min. antal deltagere:
Forventet antal deltagere:13 
Maks. antal deltagere:50 
Formelle forudsætninger:Image analysis (may be followed simultaneously)

That is,

- explain the basic concepts within image procesing.

- describe the related methods and algorithms.

- develop minor image processing systems.

- implement the systems in the language Matlab.

- do basic programing in the language Matlab.

- construct larger image processing systems by concatenating own and build-in matlab routines.

Læringsmål:The student will become capable of writing plug-inns for a professional post production station \"Flint\" in c++.

The student will have got insight into modern movie restoration methodology.

The student will know state of the art tracking.

The student will be able to program small restoration and tracking algorithms in c++ and Matlab.
Fagligt indhold:The bayesian paradigm for computer vision.

Image information is incomplete for solving the general questions in computer vision such as restoration, tracking, object recognition. The bayesian paradigm phrases this in a probabilistic setting so as to infer optimal solution.


When movies have been damaged, restoration may be performed by hand painting frame by frame. Modern digital techniques insted automated fill in the blancs by modelling the most likely contents in damaged region. We start by simple linear filtering techniques and end up with modern Partial Differential Equation (PDE)-based approaches.


In order to make special effects, artificail objects or visual pattern must follow scene structures over time, to appear in a constant position within the scene. The course goes through modern techniques starting from correlation and ending with condensation or particle filtering techniques.


The student will write computer vision algorithms in matlab and in c++ to create plug ins for the Discreete Logic Flint station, a professional digital post production facility.


The course goes through the necessary probability theory, statistics, functionals, Euler-lagrange techniques. 

Lectures in the morning. Exercises during afternoon.

Mandatory weekly exercises to be handed in. At the end of the course several weekly exercises will build towards an application. 

Eksamensform og -beskrivelse:X. experimental examination form (7-scale; external exam), 13-skala, Ekstern censur

The examination will be based on the exercises made during the course.  

Litteratur udover forskningsartikler:Handouts in terms of notes, scientific papers, and parts of books.