With this tool you can apply a trained segmentation model to an image or data set. To apply the model, Docker Desktop and a corresponding segmentation container need to be installed and running.

Parameter

Description

Model

Selects the trained model.

Output Format

When applying the processing to an image you always get an output image.

The following output formats for the processed image are available.

-

Multi-Channel

If selected, the output image will be a multi-channel image.

Each class which is defined within the model will result in a separate channel.

-

Labels

If selected, the output image will be a single-channel image in which a dedicated pixel value or color is attributed to each class.

-

Minimum Confidence

Hides all segmented pixels where the confidence value is below the threshold set by the slider.