Preprocessing

This section describes the data processing functionalities available in PlantSeg Interactive. This functionalities are available in the Preprocessing tab in the PlantSeg Interactive GUI.

Widget: Crop

To crop an image, first add a Shapes layer. You find the Add Shapes Layer button above the layer list on the left side. With the Shapes layer selected, you can add one rectangle using by pressing R and drawing.

In the Crop widget, you can also crop on the z dimension if applicable.

• Crop shapes layer: This must be a shape layer with a rectangle overlaying the area to crop.
• Z slices [start, end): Number of z slices to take next to the current selection. START is included, END is not.

Widget: Image Rescaling

Many boundary prediction models are sensitive to the scale of the input. Use the To model voxel size option to match your input to your desired model.

From factorTo layer voxel sizeTo layer shapeTo model voxel sizeTo voxel sizeTo shapeSet voxel size

Using the From factor mode, the user can rescale the image by a multiplicate factor. For example, if the image has a shape (10, 10, 10) and the user wants to rescale it by a factor of (2, 2, 2), the new size will be (20, 20, 20).

Rescale an image or label layer.

Using the To layer voxel size mode, the user can rescale the image to the voxel size of a specific layer. For example, if two images are loaded in the viewer, one with a voxel size of (0.1, 0.1, 0.1)um and the other with a voxel size of (0.1, 0.05, 0.05)um, the user can rescale the first image to the voxel size of the second image.

Rescale an image or label layer.

Using the To layer shape mode, the user can rescale the image to the shape of a specific layer. For example, if two images are loaded in the viewer, one with a shape (10, 10, 10) and the other with a shape (20, 20, 20), the user can rescale the first image to the shape of the second image.

Rescale an image or label layer.

Using the To model voxel size mode, the user can rescale the image to the voxel size of the model. For example, if the model has been trained with data at voxel size of (0.1, 0.1, 0.1)um, the user can rescale the image to this voxel size.

Rescale an image or label layer.

Using the To voxel size mode, the user can rescale the image to a specific voxel size.

Rescale an image or label layer.

Using the To shape mode, the user can rescale the image to a specific shape.

Rescale an image or label layer.

Using the Set voxel size mode, the user can set the voxel size of the image to a specific value. This only changes the metadata of the image and does not rescale the image.

Rescale an image or label layer.

Widget: Gaussian Smoothing

Apply Gaussian smoothing to an image layer.

• Sigma: Define the size of the gaussian smoothing kernel. The larger the more blurred will be the output image.

Widget: Image Pair Operations

Apply an operation to two image layers.

• Layer 1: First image to apply the operation.
• Layer 2: Second image to apply the operation.
• Operation: None
• Normalize input: Normalize the input images to the range [0, 1].
• Clip output: Clip the output to the range [0, 1].
• Normalize output: Normalize the output image to the range [0, 1].