I'm loading TIFF files, which may be a problem but I'm not sure. I can add the files via the Add Volume option, and I can play around with Thresholding etc. but when I try and load the scene again to create the 3D surface model Slicer tells me that I'm working with empty datasets.
Any ideas would be greatly appreciated!
Department of Civil Engineering
National University of Ireland (Cork)
Hi Mitch, Try saving your data in nrrd format. In the "Save" window, change the format to nrrd - but, if you have multiple series, make sure each volume has a different name so that they aren't overwritten. nrrd should be very reliable for re-opening scenes.
I'm loading TIFF files, which may be a problem but I'm not sure. I can add
the files via the Add Volume option, and I can play around with Thresholding
etc. but when I try and load the scene again to create the 3D surface model
Slicer tells me that I'm working with empty datasets.
In the Model Maker there's a Debug flag under the Debug section, try
running it with that to see what the full output of the process is (look
in the Error Log window). You do need to set the output Models before
running it. See here for more documentation:
http://www.slicer.org/slicerWiki/index.php/Modules:ModelMaker-Documentation-3.6 The slice windows can have the volume recentred by using the Fit to
Window button, the Model Maker output isn't a volume though.
In general tiff should work, yes (like .jpg or .png) as long as you
convert to single component. But there won't be any pixel or slice
spacing information. For that you need to edit the values in the
Volumes->Info pane to whatever the physical values are.
I believe Steve was referring to making the image a grayscale (single component per pixel) instead of a color image (three components per pixel). Many of the algorithms in Slicer are expecting single component images. ImageMagick tools, like convert can autoprocess a stack of TIFFs to convert from color to grayscale, if this is necessary for your data. As you've found, Slicer can read all the files in a single directory in as a single volume. However, very little is known about the volume since there was no metadata.
Please forgive me, if this is review, but the measurements, such as pixel size in each of the 2 dimensions, and the spacing in the third dimension between the slices, should be added using Slicer's Volume Module tab before making a label map. The label map process of masking what regions of the images are inside your model, can be done by hand in Slicer's editor or by running one of Slicer's Segmentation algorithms. You might want to try the GrowCut algorithm in the Editor, the Robust Statistics Segmentation, or the Fast Marching Segmentation to make your label map.
Once a labelmap is constructed, it will have the dimensions extracted from the metadata added in the Volume Module. Then running the ModelMaker should create a model.
If you are using the Grayscale ModelMaker instead to run directly off your source imagery, eliminating the explicit labelmap step, you should be able to do this once your original imagery is single component.
I know these comments are a bit sparse, but I hope it helps and makes sense when you take it along with the openpaleo tutorials you have been through. Slicer's own tutorials offer good guidance on how to repeat this process, as well.