[vtkusers] Dose Volume Histogram

Thu Oct 14 15:38:05 EDT 2010

You would have to check the neighborhood of a voxel to determine if they are inside or outside.  If your current voxel is inside, but 1 or more neighbors are outside, it’s an edge.  It probably makes more sense to build up a mask of the weighting of the voxels.  That way you only need to perform the inside/outside queries once.  From then on you can compute as needed.

                                -- Scott

From: Jothy [mailto:jothybasu at gmail.com] 
Sent: Thursday, October 14, 2010 2:23 PM
To: Scott Johnson
Cc: Xiaofeng Z; VTK
Subject: Re: [vtkusers] Dose Volume Histogram

Weighting of the voxels seems to be a good idea, but I am not sure whether you will be able to know the voxels on surface. Most of the methods says only 

whether they are inside or outside.

Jothy 

On 14-Oct-2010, at 7:44 PM, "Scott Johnson" <Scott.Johnson at neuwave.com> wrote:

	Hello Xiaofeng,

	You could minimize the partial volume by sampling into the dose grid at the resolution of the CT voxels.  For example, you iterate through the CT voxels for which you wish to create your DVH.  Map  the center of that CT voxel position and sample into the Dose Grid to find the dose value at that same position.  Then assume that dose value occupies the entire CT voxel space.  That value goes into your DVH.

	You can further minimize the partial volume effect by weighting the voxels at the edges of your ROI.  As a swag, multiply the voxel weight of an edge voxel by 0.5 .

	Other than that you need to resample your dose grid at a finer resolution still.  There is a computational tradeoff.

	Hope that helps.

	                                -- Scott

	From: vtkusers-bounces at vtk.org [mailto:vtkusers-bounces at vtk.org] On Behalf Of Xiaofeng Z
	Sent: Thursday, October 14, 2010 1:10 PM
	To: jothybasu at gmail.com
	Cc: VTK
	Subject: Re: [vtkusers] Dose Volume Histogram

	I mean voxels that's partly inside and partly outside of the polygon.

	Xiaofeng Z

________________________________

	CC: karthik.krishnan at kitware.com; vtkusers at vtk.org
	From: jothybasu at gmail.com
	Subject: Re: [vtkusers] Dose Volume Histogram
	Date: Thu, 14 Oct 2010 18:27:23 +0100
	To: xf10036 at hotmail.com

	What do you mean by partial volume?

	Jothy

	On 14-Oct-2010, at 5:48 PM, Xiaofeng Z <xf10036 at hotmail.com> wrote:

		Do you take into account of partial volume?  The dose grid is typically very coarse, I don't think the partial volume can be ignored.

		Xiaofeng Z

________________________________

		CC: karthik.krishnan at kitware.com; vtkusers at vtk.org
		From: jothybasu at gmail.com
		Subject: Re: [vtkusers] Dose Volume Histogram
		Date: Thu, 14 Oct 2010 17:23:47 +0100
		To: xf10036 at hotmail.com

		If you are speaking from the radiotherapy point of view,it should be done in a different way from what krishnan told. You should first use point-in-polygon method(which is available in vtkPolygon class) to find the voxels inside the structure and using the voxel indices, you could get their respective dose values and use a binning/histogram to bin the dose values to get the dvh along with their volumes.

		Jothy

		On 14-Oct-2010, at 5:03 PM, Xiaofeng Z <xf10036 at hotmail.com> wrote:

			Hi Karthik,  Thanks for your answer.  The problem is that I also have a contour and only the volume (area) inside the contour should be included the histogram.

			Xiaofeng Z

________________________________

			Date: Thu, 14 Oct 2010 20:14:13 +0530
			Subject: Re: [vtkusers] Dose Volume Histogram
			From: karthik.krishnan at kitware.com
			To: xf10036 at hotmail.com
			CC: vtkusers at vtk.org

			No there isn't, but it should be relatively straightforward. On your dose image,  run vtkContourFilter with isovalues, that evenly split your scalar range into the desired number of bins. Then use vtkMassProperties to compute the area within each dose isocontour and use the charting capability to plot.

			On Thu, Oct 14, 2010 at 7:52 PM, Xiaofeng Z <xf10036 at hotmail.com> wrote:

			Hello VTK users,

			Is there a function to generate Dose Volume Histogram?

			A Dose Volume Histogram is, in essence, a histogram of the area vs. intensity within an arbitrary contour i.e. the contour points are not always at pixel center.

			Thanks!

			Xiaofeng Z

			_______________________________________________
			Powered by www.kitware.com

			Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html

			Please keep messages on-topic and check the VTK FAQ at: http://www.vtk.org/Wiki/VTK_FAQ

			Follow this link to subscribe/unsubscribe:
			http://www.vtk.org/mailman/listinfo/vtkusers

			_______________________________________________
			Powered by www.kitware.com

			Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html

			Please keep messages on-topic and check the VTK FAQ at: http://www.vtk.org/Wiki/VTK_FAQ

			Follow this link to subscribe/unsubscribe:
			http://www.vtk.org/mailman/listinfo/vtkusers

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20101014/a3bd6ee8/attachment.htm>