<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<br>
<div class="moz-cite-prefix">On 13/06/2016 14:00, David Gobbi wrote:<br>
</div>
<blockquote
cite="mid:CANwS1=E3ni1+v6SPayGPyUU8CX0-Da08_X3LGW8AX_-BK1LMQQ@mail.gmail.com"
type="cite">
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote">On Mon, Jun 13, 2016 at 4:17 AM,
Martin Genet <span dir="ltr"><<a moz-do-not-send="true"
href="mailto:martin.genet@polytechnique.edu"
target="_blank">martin.genet@polytechnique.edu</a>></span>
wrote:
<blockquote class="gmail_quote" style="margin:0px 0px 0px
0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">
<div text="#000000" bgcolor="#FFFFFF"><br>
I was wondering: has there been any progress on this
side? It would<br>
quite useful. Any thought on the implementation? Thanks!</div>
</blockquote>
<div><br>
</div>
<div>No progress, it's still just an idea. When I wrote
that wiki page, I was thinking the gradient would be
useful if these interpolators were used for either
isosurfacing or ray casting. What were you thinking of
using the gradient for?</div>
<div><br>
</div>
<div>The computation of derivatives for b-spline
interpolation is currently implemented in the
vtkBSplineTransform class, which uses the derivatives to
provide the Jacobian matrix for the transformation:</div>
<div><a moz-do-not-send="true"
href="https://gitlab.kitware.com/vtk/vtk/blob/master/Filters/Hybrid/vtkBSplineTransform.cxx#L202">https://gitlab.kitware.com/vtk/vtk/blob/master/Filters/Hybrid/vtkBSplineTransform.cxx#L202</a><br>
</div>
<div>Computing a gradient would be similar, but would need
only three derivatives instead of nine.</div>
<div><br>
</div>
</div>
</div>
</div>
</blockquote>
<br>
Right now I'm working on some (finite element-based) image
correlation tool, for which I need to compute some integral on the
image (the residual), as well as some integral on the image gradient
(the search direction).<br>
<br>
For the moment I'm using an ImageInterpolator for the images, and an
ImageGradient+ImageInterpolator for their gradients, but of course
that way the interpolated gradients are not the actual gradients of
the interpolated images.<br>
<br>
Martin<br>
<br>
</body>
</html>