Skip to content

Blur Polygons

Overview ​

Applies Gaussian blur to the of each polygon in an image. The entire rectangular region around the polygon is blurred, not just the polygon interior.

Use Blur Polygons when:

  • You need to obscure specific areas detected by a model.
  • You want to hide personal data for privacy compliance.

💡 Tip

Higher Sigma values result in a stronger blur, effectively obscuring more detail. The blur kernel size scales automatically with the image resolution — larger images produce a proportionally larger blur kernel.

â„šī¸ Note

Each polygon is converted to its bounding box before blurring. The entire rectangular region is blurred, not just the area inside the polygon shape.

Inputs ​

NameTypeDescription
input_imageimageThe source image to apply the blur on.
polygon_listlist[polygon]The list of polygons defining the regions to be blurred.

Outputs ​

NameTypeDescription
blurred_imageimageThe resulting image with the blurred regions.

Parameters ​

ParameterDefaultValid valuesTunable
Sigma151 – 1000❌

Sigma ​

Specifies the strength of the Gaussian blur.

Example ​

Obscure detected PCBs on an assembly line ​

A model detects two PCB boards on an assembly line and outputs their outlines as polygons. The Blur Polygons node blurs the bounding-box rectangle around each detection to hide proprietary circuit layouts.

Configuration:

  • Sigma: 25
Input image with two PCB detections highlighted by green polygons
Input — two PCBs detected as polygons.
Output image with the bounding-box regions around each PCB blurred
Bounding-box regions blurred with Sigma = 25.

Result: Each polygon's bounding box blurred, hiding the circuit detail while leaving the rest of the image untouched.

Reference: