CVPR Workshop on Registration of Very Large Images

June 23, 2014

Columbus, OH, USA
Description
Throughout the years elaborate and robust methods have been developed for the registration of various types of images. Such
methods often become impractical when applied to very large images, typically those containing 100M or more pixels/voxels.
The objective of this workshop is to bring together modified past methods as well as new methods that are particularly
effective in the registration of very large images. In addition to the traditional image registration topics, this workshop covers
template matching, i.e., finding objects of interest (templates) in a very large image or in a very long video, and registering
very large range images containing 100M or more points.
Topics
  • Fast similarity and distance measures
  • Efficient point and line detectors and descriptors
  • Efficient point pattern matching algorithms
  • Fast RANSAC and other robust correspondence algorithms
  • Nonrigid transformation functions for alignment of very large images
  • Mosaicking of a large number of (>100) overlapping images
  • Methods for registration of multivariate/hyperspectral images
  • Methods for registration of 4-D medical images
  • Methods for registration of electron microscopy images
  • Metrics for evaluation of registration accuracy of very large images
  • 3-D scene reconstruction from registration of a long video sequence
  • Methods for registration of very large range images
  • Methods for detecting objects of interest in a very large image or in a very long video
Important Dates
  • New submission deadline: 5:00 PM ET, March 21, 2014
  • Notification of acceptance: 5:00 PM ET, April 17, 2014
  • Papers in final form: 5:00 PM ET, May 2, 2014
Paper Submission
Papers should be prepared according to CVPR guidelines. To submit a paper to this workshop, please visit CMT site
https://cmt.research.microsoft.com/IRW2014/   and upload your paper. Reviewing will be double-blind; therefore, information
that can identify an author should not be included in the submitted paper.  
Organizers
  • Ardy Goshtasby, Wright State University, USA,  E-mail: ardy at wright.edu (general)
  • Chang Shu, National Research Council, Canada, E-mail: chang.shu at nrc-cnrc.ca (range)
  • Akihiro Sugimoto, National Institute of Informatics, Japan, E-mail: sugimoto at nii.ac.jp (range)
  • John Camp, Air Force Research Lab, USA, E-mail: john.camp.1 at us.af.mil (range)
  • Nathan Netanyahu, Bar-Ilan University, Israel, E-mail: nathan at cs.biu.ac.il (remote sensing)
  • Claude Cariou, University of Rennes, France, E-mail: cariou at univ-rennes1.fr (remote sensing)
  • Hector Erives, New Mexico Tech., USA, E-mail: erives at tee.nmt.edu (hyperspectral)
  • Clark Taylor, Air Force Research Lab, USA, E-mail: clark.taylor.3 at us.af.mil  (aerial)
  • Lyubomir Zagorchev, Philips Research, USA, E-mail: lyubomir.zagorchev at philips.com (medical)
  • Marcel Jackowski, Univ. Sao Paulo, Brazil, E-mail: mjack at ime.usp.br (medical)
  • Martin Satter, Kettering Medical Center, USA, E-mail: martin.satter at khnetwork.org (medical)
  • Simon Warfield, Harvard, USA, E-mail: simon.warfield at childrens.harvard.edu  (electron microscopy)

For further information, please contact the organizer in your
area of interest.

Sponsor: Image Registration and Fusion Systems