Threshold

The Threshold command is the primary segmentation step. It converts a greyscale image to a binary mask: pixels with intensity below the minimum threshold become 0 (background); pixels above become 65535 (foreground).

The output of Threshold feeds directly into Connected Components.

Single threshold

Define one threshold entry to produce a simple binary mask:

Parameter	Description
Method	Manual or one of the auto-threshold algorithms
Min threshold	Minimum intensity to include as foreground
Max threshold	Maximum intensity to include as foreground (set to 65535 to disable)
Unit	Intensity unit (Absolute 0–65535, Percent 0–100, or Relative 0–1)
Object class	The segmentation class assigned to objects detected by this threshold entry

Tip

Examine the image histogram to choose a starting value for the minimum threshold. Set it just above the highest background pixel intensity.

Caution

Always set a minimum threshold greater than zero even when using an auto-threshold method. Control images containing only background noise will otherwise produce false-positive detections.

Multiple threshold classes

Click + Add to define additional threshold entries for the same image. Each entry uses a different intensity range and assigns a different segmentation class, allowing two or more object populations to be detected from one image in a single step.

Under the hood, each threshold class maps to a distinct greyscale value in the binary output (65535 for the first class, decrementing for subsequent classes). Connected Components and Extract ROIs use these values to distinguish the populations.

Auto-threshold algorithms

When Method is not Manual, EVAnalyzer analyses the image histogram and computes a threshold automatically. Available methods include Otsu, Triangle, Mean, and others.

Auto-thresholds can fail on empty images (no objects, only noise) because they always find a separation. Prevent this by setting a non-zero Min threshold as a lower bound: if the computed value falls below it, EVAnalyzer uses the minimum instead.

Maximum threshold

Setting Max threshold to a value below 65535 restricts detection to pixels within the [min, max] range. This allows extraction of a specific intensity band — for example to separate bright artefacts from dim objects, or to extract the background itself by setting Min to 0 and Max to just below the signal level.