Select Import → ImportGeo-Vol in the menu bar to convert volume and image formats, such as µCT-scans, into 3D material models. The ImportGeo-Vol section opens at the left side of the GeoDict GUI.

Click Browse to open the Select File dialog. Browse for a file or an image stack. Select one file and click Open.

After clicking Open, the Geometry Import dialog opens. Depending on the file format, a dialog opens containing various options and parameters in preparation for the import. For most formats, the size of the file’s structure is automatically entered. The amount of approximately required memory and free memory are displayed and indicate the amount of working memory available.

NX, NY, and NZ specify the size of the structure in voxels. For most file types, these values are entered automatically. For RAW format files, you need to know these values and enter them here.

The double sliders allow restricting the portion of the 3D image to be loaded. The left-hand sliders define the number of voxels which are cut away at the start of the image in each direction, and the right-hand sliders define the last voxel to be imported in each direction.

By default, ImportGeo-Vol imports the complete image. The values for NX, NY, and NZ, and the first and last voxel positions can be changed for very large images that would not fit in memory. It is possible to focus on any sub-cuboid of the original image by adjusting the sliders.

Below the the size of the file’s structure and the cropping options, define the Voxel Length in the selected unit.

Check Keep existing Volume Fields, if the loaded volume fields should not be discarded. This is useful, if more than one image dataset should be used for AI-Segmentation training. If the scans don’t have the same dimensions, a dialog opens after clicking Image Processing, asking if the structure and volume fields should be discarded. When clicking Yes, the domain size of the first image is kept in the visualization area, which does not fit to the new image.

For most importable file formats, the size parameters are automatically entered after browsing for the file and loading it. Click Quick Segmentation to bring in the structure saved in the selected file, so that it is displayed in the visualization area. All supported file formats can be processed (i.e., the objects in the structure can be cut, resized, rotated, filtered, and segmented) before the file is finally imported into GeoDict.

At the bottom of the dialog, two options for continuing are available:

• Quick Segmentation: For many of the importable file formats, the structure is directly loaded after clicking Quick Segmentation.
  For other files, such as *.raw image files, or stacks of *.png files, that are considered a three-dimensional image with gray values (e.g., from 0 to 255), clicking Quick Segmentation opens the 3D Image Quick Segmentation dialog. This dialog shows a histogram with the number of voxels for each of these gray values and enables a simple global segmentation without further 3D Image Processing.
• Image Processing: Allows for more complex image processing, e.g., changing image sizes or orientation, filtering the images, and finally creating a segmentation.