Thresholding Tutorial
Setting a proper threshold is one of the most important steps when segmenting a model that accurately reflects the anatomy of interest. The Threshold Scroll Bar in Elucis allows you to define values that correspond to the specific tissue densities you wish to include in your segmentation.
1. Preparing to Set a Threshold
Open the Controller Menu.
Click Models, then create a New Structure.
On the Stylus Menu, select the Add button.
At the bottom left of the Surface Plane, select Display Threshold.
You’ll now see regions highlighted in pink — these areas represent your current threshold range, which can be modified using the scroll bar in the center of the surface plane.
2. Understanding the Threshold Scroll Bar
The scroll bar has two slider buttons — a left and a right — that define the minimum and maximum range of your threshold:
Moving the sliders closer together creates a narrow threshold range, capturing a smaller density window.
Moving them further apart creates a wider threshold range, encompassing more densities.
3. General Threshold Ranges (CT Data)
When working with CT image sets, the scroll bar roughly corresponds to tissue density in Hounsfield units (HU):
Left side → Radiolucent structures (air, fat, low density).
Middle → Soft tissue (muscle, organs).
Right side → Radiodense tissue (bone, contrast media, metal).
Example:
If segmenting the myocardium while excluding the contrast within heart chambers:
Move the left slider until the myocardium appears highlighted.
Move the right slider until the contrast regions are no longer highlighted.
The threshold range will narrow, indicating that only the densities within that range are included.
4. Thresholding in MR and Echo Data
For MRI and Echocardiography datasets:
Use the left slider to broadly capture the tissue of interest.
Adjust the right slider to refine and isolate the desired region.
This process allows flexible adaptation to different intensity scales used in non-CT modalities.
5. Fine-Tuning Threshold Values
For greater precision:
Gently drag the slider buttons for small adjustments, or
Use the Minimum and Maximum toggles located on the right-hand side of the surface plane.
These fine controls allow you to capture subtle tissue boundaries accurately.
6. Reducing Noise
If the threshold appears noisy or contains unwanted speckles:
Check the Reduce Noise box.
You’ll see that the threshold becomes smoother, with cleaner, more defined borders.
Note that noise reduction applies only to the threshold visualization, not the image data itself.
By mastering thresholding, you ensure that your segmentations closely match true anatomical boundaries — providing a clean, accurate foundation for further refinement and analysis.