Call for Papers: IEEE Workshop on Three Dimensional Information Extraction
for Video Analysis and Mining (in conjunction with CVPR 2010)

**Important dates**:
Deadline: March 14
Notification: April 4
Camera-ready: April 11
Workshop: June 14

For more details, please see our website at:
http://ww2.cs.fsu.edu/~donate/public/workshop/

With the availability of large video databases, video analysis and
mining has become an active of research involving the computer vision,
image processing, and information retrieval communities. Many
algorithms have been proposed and numerous systems have been
implemented, most of which are based on extrinsic representations of
scenes by extracting and utilizing image features from video frames.

The goal of this line of research is to extract the three dimensional
structure of the scene observed in a video, and use it to improve the
performance on various video analysis tasks. Such tasks may include
shot-boundary detection, scene segmentation, video summarization,
content-based video retrieval, and so on. Since video frames are two
dimensional projections of three dimensional scenes, intrinsic three
dimensional representations (as well as associated methods) should
provide more robust and meaningful information to improve the
performance of these video analysis tasks.

Intrinsic three dimensional information in videos, however, has not been
widely employed. Extracting three dimensional information is not only a
computationally intensive task, but the problem contains many other
inherent challenges. Such challenges include dealing with 3D information
extraction from a single view of a scene (as opposed to using a stereo
camera), unknown camera calibration information of a scene, designing
methods robust enough to deal with arbitrary object and camera motions,
etc.

The goals of the workshop are mainly two folds. First, to promote the
computer vision community to adopt its expertise on video analysis
tasks; these tasks involve all core image analysis and computer vision
problems, and provide a good testbed for testing algorithms. Second, to
act as a bridge among several communities to facilitate collaborations
and cross-cutting research topics that are mutually beneficial.

Papers describing novel and original research are solicited in the areas
related to video analysis and 3D information extraction. Some topics
include (but are certainly not limited to):

? New 3D information extraction techniques for video analysis.
? Mathematical models of 3D objects for video analysis
? Efficient and effective representations of 3D video data for
retrieving and indexing.
? Video analysis and mining systems, and applications that rely on 3D
information
? Methods for extracting 3D information from videos.
? 3D feature descriptors from 2D video data.
? Camera localization from videos.
? Content-based video searching/retrieving using 3D data.
? Shot/scene boundary detection incorporating intrinsic 3D information
? 3D motion analysis in videos.

Organizing Committee:
? Bir Bhanu, University of California, Riverside
? Arturo Donate, Florida State University
? Xiuwen Liu, Florida State University
? Igor Kozintsev, Intel Labs, Intel Corporation

Program Committee:
? J. K. Aggarwal, The University of Texas at Austin
? Bir Bhanu, University of California, Riverside
? Andrew Davison, Imperial College London
? Arturo Donate, Florida State University
? Richard Hartley, Australian National University
? Omar Javed, ObjectVideo Inc.
? Rangachar Kasturi, University of South Florida
? Igor Kozintsev, Intel Labs, Intel Corporation
? Xiuwen Liu, Florida State University
? Yaser Sheikh, Carnegie Mellon University
? Dan Schonfeld, University of Illinois at Chicago
? Dacheng Tao, Nanyang Technological University
? George Toderici, Google, Inc.