Work on using the projective group model
for image mosaicing was previously presented:

<ul>

<li>First presented in 1993, where the parameters of a
homographic (8-parameter) coordinate transformation
were adjusted for each image to minimize the mean squared difference
(or maximize the inner product) between that image and a
selected reference image.
<pre>
S. Mann
Compositing Multiple Pictures of the Same Scene
The Society of Imaging Science and Technology
Proceedings of the 46th Annual IS&T Conference
Cambridge, Massachusetts
May 9-14
YEAR 1993
  </pre>

<li>Virtual Bellows:  Constructing High Quality Stills from Video
   <p>
   IEEE First Int. Conf. on Image Proc, Austin, Texas, 1994
   <p>
   You can look at the
   <a href="http:/vismod/publications/tech_reports/abstracts/TR-259-ABSTRACT.html">abstract</a>
   or retrieve a
   <a href="http:/vismod/publications/techdir/TR-259.ps.Z">PostScript file</a>.

<p>
<li>The above work is explained in more detail in an internal (unpublished)
  Technical Report:
  <pre>
TR-278
S. Mann and R. W. Picard
New Perspectives on the Motion Description of a Rigid Planar Patch,
With Application to Hi-res Stills from Video:
Generalized Large-Displacement 8-Parameter Motion
M.I.T. Media Lab Perceptual Computing Section
March 10, 1994
  </pre>
  You can look at the
   <a href="http:/vismod/publications/tech_reports/abstracts/TR-278-ABSTRACT.html">abstract</a>
   or retrieve a
   <a href="http:/vismod/publications/techdir/TR-278.ps.Z">PostScript file</a>.

<p>
<li> Richard Szeliski has also presented a similar result:
  <pre>
Richard Szeliski and James Coughlan
Hierarchical Spline-Based Image Registration
CVPR, Seattle, Washington, 1994
  </pre>
</ul>

<P>
Return to <a href="claire.html">ICCV workshop proposal</a>
<P>
Return to <a href="/~steve/steve.html">Steve's home page</a>