CFP – First International Workshop on Parts and Attributes (ECCV)

First International Workshop on Parts and Attributes
In conjunction with ECCV 2010, September 10th, 2010, Crete, Greece

Important Dates:
* Deadline for submission of papers: June 16th, 2010
* Notification of acceptance: July 6th, 2010
* Final version of submission: July 14th, 2010

Recent advances in probabilistic modeling and optimization have lead to
a renewed interest in part-based methods for solving fundamental
problems in computer vision, in particular object detection,
classification, and pose estimation. Although substantial progress has
been made to adequately parse objects into parts and build models to
handle variations such as object pose and lighting, many open problems
still remain to be solved. At the same time, part-based models for scene
understanding are being developed that for the first time allow a
holistic understanding of natural scenes based on local image regions
and their interaction. In parallel to part-based models, attributes
based classification has recently been rediscovered, e.g. as a promising
tool to overcome the problem that there are too many visual object
categories to train individual classifiers for each of them. Attributes
also make a more targeted search in large image databases possible,
allowing e.g. queries like “person with blond hair, long nose and a red
shirt”. The goal of this workshop is to bring together emerging research
on part-based methods and attributebased methods. We believe there is a
strong link between these two research areas which has not been
previously explored.

Call for Participation:
We invite high quality, original submissions for oral presentation
during the workshop. Contributions
from the following areas are especially welcome:

* Part-based Methods:
– Localization of object parts
– Deformable and rigid part-based models
– Generative vs. discriminative part-based models
– Structured prediction for part estimation
– Context and hierarchy in part-based models

* Attribute-based Methods:
– Learning visual attributes across object classes
– Attribute-based classification with few examples
– Semantic attributes as object representations

* Hybrid Part/Attribute-based Methods
– Semantic parsing of objects and scenes
– Joint learning of object parts and attributes

* Applications:
– Object detection and recognition
– Visual image search
– Soft-biometrics
– Innovative applications of parts and attributes

We also invite submissions from related domains including theoretic
results relevant to the workshop’s topic. Papers must be in PDF format
and should not exceed 14 pages (ECCV format). All submissions are
subject to a double-blind review process by the program committee.
Further details can be found on the workshop homepage.

Workshop Chairs:
– Rogerio S. Feris, IBM
– Tiberio Caetano, NICTA
– Christoph H. Lampert, IST Austria
– David Forsyth, UIUC