home
research
publications
teaching
|
Basic
Research
| Applied Research (Projects) |
Open
Positions for PhD Students
Basic Research:
| Title |
Classification
of visual data in high dimensional spaces
|
 |
| Members: |
Dr.
Jordi Vitrià, Marco Bressan, David Guillamet, David Masip.
|
|
|
| Description: |
High
dimensional data appears in many pattern
recognition problems such as remote
sensing, appearance-based object recognition,
text categorization, etc. A stochastic
approach for the classification of high
dimensional data is always a delicate
issue. For linear or quadratic classifiers
the number of training samples depends
linearly or quadratically on the data
dimensionality subset of features, an
exhaustive sequential feature selection
procedure is required, so the size of
the problem grows combinatorially on
the dimension. Furthermore, the training
sample size needs to increase exponentially
in order to effectively estimate the
multivariate densities needed to perform
nonparametric classification. To avoid
the problem of dimensionality, the most
common approach is the implementation
of feature extraction or dimensionality
reduction algorithms. Our approach does
not seek dimensionality reduction followed
by the implementation of parametric
or nonparametric techniques for density
estimation. Instead, it focuses on the
higher level statistical properties
of the data (independent component analysis,
non negative matrix factorization),
which is transformed in such a way that
density estimation in the transformed
space is simplified and more accurate.
We are also working on the use of Naïve
Bayes Classifier under this approach.
|
| Keywords: |
Independent
Component Analysis, Non-negative Matrix
Factorization, Naïve Bayes Classifier.
|
|
| Literature: |
M.Bressan,
D.Guillamet, J.Vitrià. Using an ICA
Representation of Local Color Histograms
for Object Recognition. Pattern Recognition
(accepted for publication). April 2002.
M.Bressan, D.Guillamet, J.Vitrià.
Multiclass Object Recognition using Class-Conditional
Independent Component Analysis. Submitted
for publication to Cybernetics and Systems.
March 2002.
M.Bressan, J.Vitrià. Independent
Feature Selection. Submitted for publication
to IEEE Trans on PAMI. December 2001.
|
|
|
|
Applied Research (Projects):
 |
|
|
| Contact: |
Dr.
Jordi Vitrià
Computer Vision Center
Edifici O, Campus Universitat Autònoma
de Barcelona
08129 Bellaterra (Barcelona) Spain
Phone: +34 93 581 18 28
E-mail: jordi@cvc.uab.es
|
|
|
| Description: |
CorkInspect
is an IST EU project within the EUTIST-IMV
cluster.
Corkinspect is
a Trial project intending to support,
as a major objective, the early customisation
and validation of a Computer Vision
solution for cork stopper production
control. The main discriminant feature
of CorkInspect is its innovative development
based on its way of building inspection
and control rules. CorkInspect establishes
its control rules
through an iterative procedure based
on statistical control features and
boundaries tolerances from selected
samples. That brings great flexibility
and, in special, when we work with natural
products, this flexibility to establish
control features becomes a must.
|
| Time Frame: |
January 2002-March 2003 |
|
| Literature: |
Petia Radeva(CVC)
, Marco Bressan(CVC), Antonio Tobar(CVC)
and Jordi Vitrià(CVC). Real Time
Inspection of Cork Stoppers using Parametric
Methods in High Dimensional Spaces.
IASTED International Conference, Signal
and Image Processing (SIP 2002), August
12-14, 2002, USA.
David Guillamet (CVC), Jordi Vitria(CVC).
Determining a Suitable Metric When
using Non-negative Matrix Factorization
David Guillamet (CVC), Jordi Vitria(CVC).
International Conference on Pattern Recognition
(IAPR 2002), August 11-15 2002 - Québec
City Convention Center.
|
|
|
|
|
Augmented
Reality and Ambient Intelligence
|
|
| Contact: |
Dr.
Jordi Vitrià
Computer Vision Center
Edifici O, Campus Universitat Autònoma
de Barcelona
08129 Bellaterra (Barcelona) Spain
Phone: +34 93 581 18 28
E-mail: jordi@cvc.uab.es
|
|
|
| Description: |
The
perception of reality can be enhanced
making connections between physical
and virtual world. Computer vision is
a candidate to implement this connection.
Example I:
Visual Tags. Paper is a technology
that will continue to be the ideal
choice for certain activities. Emphasis
has to be put on the development of
technologies which attemp to better
integrate this paper use with coexisting
digital technologies.
Example II:
Information in Places. What if we
could put information in places? More
precisely, what if we could associate
information with a place or an object
and perceive the information as if
it were really there?
Ambient Intelligence
builds on three recent key technologies
(Ubiquitous Computing, Ubiquitous Communication
and Intelligent User Interfaces) for
developing the next generation of Information
Processing systems. Under this paradigm,
people will be surrounded by intelligent
and intuitive interfaces embedded in
everyday objects around us and an environment
recognizing and responding to the presence
of individuals in an invisible way.
Our environment must be aware of human
presence, personalities, needs and must
be capable of responding intelligently
to spoken or gestured indications of
desire, and even in engaging in intelligent
dialogue. Visual perception of objects
and events will play an important role
in this development.
|
| Time Frame: |
January 2002-March 2003 |
|
| Literature: |
|
|
|
|
|
The
following positions are currently available for a
PhD:
 |
Face Recognition
and Analysis for Context Aware Environments. |
|
Published:
Position: |
25.6.2002
PhD Thesis |
| Contact: |
Dr. Jordi
Vitrià
Computer Vision Center
Edifici O, Campus Universitat Autònoma
de Barcelona
08129 Bellaterra (Barcelona) Spain
Phone: +34 93 581 21 69
E-mail:jordi@cvc.uab.es
|
|
|
| Description: |
Ambient Intelligence builds on three
recent key technologies (Ubiquitous
Computing, Ubiquitous Communication
and Intelligent User Interfaces) for
developing the next generation of Information
Processing systems. Under this paradigm,
people will be surrounded by intelligent
and intuitive interfaces embedded
in everyday objects around us and an
environment recognizing and responding
to the presence of individuals in an
invisible way. Our environment must
be aware of human presence, personalities,
needs and must be capable of responding
intelligently to spoken or gestured
indications of desire, and even in engaging
in intelligent dialogue. Visual perception
of people and events will play an important
role in this development. This PhD thesis
will focus on the development of real
time face analysis system and its use
on contex aware environments.
|
| Time Frame: |
Beginning in January 2003 |
|
| Literature: |
D.Guillamet,
B.Schiele,J.Vitrià.
Analyzing Non-Negative Matrix Factorization
for Image Classification.
Accepted
to ICPR 2002.
D.Guillamet,
J.Vitrià. Non-negative matrix
factorization to extract part-based representations.
Butlletí de l'ACIA, n.25, 2001,
pp. 255-262.
A.Pujol,
J.Vitrià, F.Lumbreras, J.J.Villanueva.
Topological principal component analysis
for face encoding and recognition. Pattern
Recognition Letters, pp. 769-776, Volume
22, Issue 6-7, May 2001.
M.Bressan,
D.Guillamet, J.Vitrià.
Independent Feature Selection.
Submitted for publication to IEEE Trans
on PAMI. December 2001. |
|
| Requirements: |
We
are looking for a highly motivated student
with strong interest in computer vision
and pattern recognition, especially in
the areas of classification and Bayesian
analysis. Solid mathematical and programming
skills are also required. |
|
|
|
|