Several open research positions are available from January 2007,
in the computer vision group Psi/Visics of the University of Leuven, 
Belgium,
(http://www.esat.kuleuven.be/psi/visics/index.php), both at PhD- and 
post-doc level.

All focus on image interpretation and/or multimedia search and 
summarization,
and fall within the framework of the European Project CLASS 
(Cognitive-Level
Annotation using Latent Statistical Structure, http://class.inrialpes.fr)
or the Flemish IWT project AMASS++ (Advanced Multimedia Alignment and
Structured Summarization).

1. Realtime retrieval from large video archives
This topic includes the development of efficient, spatio-temporal local 
features and descriptors,
fast matching algorithms, as well as efficient verification schemes.
The applicant for this topic should have especially good programming 
skills.

2. Attributes, context and compositions
In a second topic, the aim is to improve the state-of-the-art in image 
understanding by learning
labels common to several object classes (attributes like 'red' or 
'big'), learning compositions of
object parts and exploiting context information.

3. Scaleability of categorization
Another researcher is to look into the problem of recognizing, 
discriminating and
localizing large numbers of classses (>100) simultaneously. This will 
include topics such
as feature sharing and building class hierarchies.

4. Multimodal news summarization
Finally, we want someone to tackle the problem of news broadcast 
summarization, by
exploiting both visual and textual information in an integrated fashion.

Leuven is a small university town in the Dutch-speaking part of Belgium, 
just 20 kms from Brussels.
Most people are fluent in English though, and English is the standard 
language in the lab.

Candidates should have a Master of Applied Science (or an equivalent 
diploma), strong
background in applied math and statistics, good programming skills, 
preferrably in C or C++ (experience
working in a linux environment is a plus), skills in English, and lots 
of enthusiasm.
For post-docs, experience in computer vision and a good track record of 
relevant publications is a must.

Applications should be sent by email (PDF-format) before December 30th, to
Tinne.Tuytelaars@esat.kuleuven.be or Luc.Vangool@esat.kuleuven.be

Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm