Q&A : 3D Photography Challenge
Vision List Digest:
Article 1,
Volume 23, Issue 6
From: "Jan Neumann"
Post-Followup: submission@VISLIST.com
In connection with the Int'l Symposium for 3D processing,Visualization
and Transmission
( http://www.umiacs.umd.edu/conferences/3dpvt04/index.htm
) that covers
topics of interest for researchers in both computer vision and computer
graphics, we will organize a "3D Photography Challenge"
( http://www.umiacs.umd.edu/conferences/3dpvt04/challenge.htm
).
The reconstruction of 3D shape and motion from video is one of the basic
problems that people in our field have worked on for a long time.
Unfortunately, it has always been difficult to compare the results of
different algorithms since there are no publicly available datasets
consisting of multiple, calibrated image sequences that everyone has
access to and could use as a benchmark data set.
Therefore, we make available three challenging datasets, so that
everyone can apply their algorithms to these sequences and the
resulting reconstructions can be compared against eachother. Each data
set consists of synchronized and calibrated image sequences taken from 33
different view positions, using the Keck laboratory at UMD
( http://degas.umiacs.umd.edu/keck/spec.html
).
For more information please see:
http://www.umiacs.umd.edu/conferences/3dpvt04/challenge.htm
.
The submitted reconstructions and/or animations will be judged by a
group of established members of our community who will award a prize to
the best work. Details about the submission process and the judging
criteria will be available on the website by mid-April.
With best regards,
Jan Neumann & Patrick Baker
Computer Vision Lab, University of Maryland
http://www.vislist.com