TU Berlin

Processing Digital Camera Images

our cg logo

Inhalt

zur Navigation

Es gibt keine deutsche Übersetzung dieser Webseite.

Processing Digital Camera Images (WS 2011/12)

0433 L 360
Instructor:
Prof. Dr. Marc Alexa, Prof. Dr. Bernd Bickel
Contact
Topics/Schedule:
Administration:

Course structure:
Seminar - 2 SWS
Room:
E-N 719
Date:
Tuesday 4pm - 6pm  
Start: 18 October 2011
End: 15 February 2012
Description:
Images taken with a digital camera can be processed and manipulated to create novel imagery unachievable with a single shot. In this course we will computationally create novel imagery by combining a large number of existing images, coming either from multiple shots of the same scene taken with varying camera parameters or from existing large image databases (e.g. Flickr).
Requirements:
Interest in visual arts/photography as well as programming skills (Matlab/C++). It is strongly recommended that students have already taken introductory Computer Graphics or Computer Vision courses.
 
Details:
The course gives a brief overview of the basic ideas and several recent developments in the field of Computational Photography as well as Data-Driven Vision and Graphics. Students will work in small teams on a chosen topic and present the basics of their topic in an intermediate talk. Finally, each team will implement a recent research paper and present their topic, application and results in a concluding talk. Both efficiency of the underlying algorithms as well as visual quality of the resulting imagery are important aspects in this course. Grading will be based on all aspects of performance.
Registration:
Please register for the course until October 25th by sending an eMail mentioning the course title (PDCI), course of studies (e.g. Informatik), your name and your student id (Matrikelnummer) to .

Slides

Schedule
Date 
Topic
18 Oct
Overview
25 Oct
Topics, forming teams
01 Nov
  • SIFT (Christian Thurow)
  • Structure from motion (Christopher Leiste)
08 Nov
  • Poisson equation (Stefan Chmiela)
  • Applied gradient domain imaging (Andreas Langenhagen)
15 Nov
  • RANSAC (Dennis Giovanni Balreira)
  • Image descriptors (Henrique Weber)
22 Nov
  • GraphCut (Alberto Sanchez)
  • Template matching (Michael Winkelmann)
29 Nov
  • Distance metrics (Damien Thome Lutz)
  • Nearest neighbor search (Christian Kohl)
06 Dec
Roadmap presentations
13 Dec
Project meetings
Christmas break
03 Jan
Project meetings
10 Jan
Project meetings
17 Jan
Project meetings
24 Jan
Project meetings
31 Jan
Final project presentations:

07 Feb
Final project presentations:

14 Feb
Final project presentations:

23 Feb
Documentation online (no class)

Sample projects from previous courses

Automatic colorization of a greyscale image (left) using statistics learned from a large image database. The colorized result is shown on the right.
Lupe
Interactive Digital Photomontage: automatically combining a set of macro photographs yields an image that is sharp everywhere.
Lupe

Samples of what students implemented in previous classes:

SS 11:

WS 10/11:

WS 09/10:

WS 08/09:

WS 07/08:


List of previous related courses:

Navigation

Direktzugang

Schnellnavigation zur Seite über Nummerneingabe