Generating Contour Meshes from Images

You can generate contour meshes that describe the contour around values in a dataset that are above the selected lower threshold. In the example below, a contour mesh was generated from a volumetric scan of a titanium implant in rabbit bone.

Contour mesh (on right) of a titanium implant in bone

Contour mesh

Contour meshes are also available for regions of interest (see Generating Contour Meshes from ROIs). When generated from this type of object, the contour will describe the labeled voxels contained within the region of interest.

The settings for generating contour meshes are available in the Generate Contour Mesh dialog, shown below. Right-click the required volumetric dataset in the Data Properties and Settings panel to access the panel.

Generate Contour Mesh panel

Generate Contour Mesh panel

The following options are available for generating contour meshes.

Contour mesh options
  Description
Object Indicates the currently selected object. If required, you can select another object in the drop-down menu to generate additional contour meshes.
Threshold Determines the threshold value that will be used to generate the contour mesh. You should note that the initial value computed is the Otsu threshold.
  • Enter a new value in Threshold edit box.
  • Adjust the threshold value with the slider.
  • Check the Use Otsu option to reset the current value and lock it at the Otsu threshold.
Sampling Determines the sampling rate that will be applied along each axis and the sampling algorithm. At a setting of 4, one of every four pixels will be sampled. Modifying the values for an axis or axes will generate a mesh representing more or less vertices. Each voxel value is analyzed when the sampling value is set at 1.

You can also select a sampling algorithm — Linear or Nearest.

Linear… Selecting Linear will result in a linear sampling algorithm being applied.

Nearest… Selecting Nearest will result in a nearest-neighbor sampling algorithm being applied.

Thresholding may not be suitable for noisy data. Random noise in the background with voxel values above the threshold will also be displayed and may obscure features of interest.