<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Tahoma
}
--></style>
</head>
<body class='hmmessage'>
I am planning to take the route scott suggested. I may add some edge-based interpolation to the slices inserted.<BR>
<BR>
One more thing I still not sure how to do is what algorithm to use to cap off the end contours, contours have no overlapping in one or both neighouring slices. If the caps are not added, the estimated volume will be consistently become smaller and smaller as the voxel size is reduced.<BR>
<BR>
So far I didn't have any luck finding a good extrapolation algorithm to estimate the size and shape of the caps. Does anyone have suggestions?<BR>
<BR>
Thanks!<BR>Xiaofeng<BR>
<BR>
<BR>
Scott Johnson <Scott.Johnson <at> neuwave.com> writes:<BR>> <BR>> You could minimize the partial volume by sampling into the dose<BR>> grid at the resolution of the CT voxels. For example, you iterate through<BR>> the CT voxels for which you wish to create your DVH. Map the center<BR>> of that CT voxel position and sample into the Dose Grid to find the dose <BR>value<BR>> at that same position. Then assume that dose value occupies the entire CT<BR>> voxel space. That value goes into your DVH.<BR>> <BR>> <BR>> <BR>> You can further minimize the partial volume effect by weighting<BR>> the voxels at the edges of your ROI. As a swag, multiply the voxel weight<BR>> of an edge voxel by 0.5 .<BR>> <BR>> <BR>> <BR>> Other than that you need to resample your dose grid at a finer<BR>> resolution still. There is a computational tradeoff.<BR>> <BR>> <BR>> <BR>> Hope that helps.<BR>> <BR>> <BR>> <BR>> --<BR>> Scott<BR>> <BR>> <BR>> <BR>
<BR><BR> <BR> </body>
</html>