3-D Registration : Theory and Applications

                      Sameh Yamany
                    U. of Louisville


                        Abstract

With the proliferation of high quality 3-D capturing sensors, especially
in the computer vision arena, and volumetric image data, especially for
the medical community, 3-D surfaces and objects become available and ready
for structural analysis. Most often, 3-D surfaces or objects are
represented by a triangulation of the surface or by a binary voxel
representation. Although these representations make possible a 3-D
rendering for visually capturing the object properties, both lack
descriptive power as they are based on huge list of surfaces elements or
voxels. Characterizing and understanding shape properties, however,
requires a representation which captures global and local shape features
with small number of parameters. Such a concise description could be
useful for comparing various objects, finding dissimilarities, recognizing
based on predefined models, registering two objects and for efficient
reconstruction and manipulation of objects. This talk introduces a
representation scheme for surface registration that has the above
features. Applications presented include object recognition, surface
matching, medical volumes registration and surface integration.


Bio

Sameh M. Yamany, obtained his Ph.D. in Computer Science and Engineering
from the University of Louisville in 1999.  He obtained his M.S. and B.Sc.
degrees from the Systems and Biomedical Eng. Dept., Cairo Unv. in 95 and
91 respectively. He worked as TA in the faculty of Eng., Cairo Unv. from
91 to 95, and at the same time, as R.A. in the Cairo Univ. Center for
Advanced Software Development and Applications (ASDA)  in the field of
natural language processing, neural computing, OCR, and multimedia. He
worked at IBM Cairo Scientific Center in 95. He is currently a research
scientist at the Computer Vision and Image Processing (CVIP) lab at the
University of Louisville. His current research interests include 3-D
visualization, computer vision, image processing, and pattern recognition.
He has assisted in three research projects funded by the NSF, DoD and NIH
institutions. He received an award from the CVIP lab on 1997 for his
research in the area of pattern recognition. He also received an award in
the 1998 from the UofL Speed School of Engineering for his research in the
area of computer vision and 3D model building. He is currently a member of
the IEEE, EMBS and SMC societies.