Crop Dataset¶

Crop the image and corresponding Selection, Mask, and Model layers.

Cropping can be done in:

Interactive
Manual
ROI mode

Demonstration

Crop dataset demonstration

Interactive mode¶

When the Interactive mode is selected, you can draw a rectangular area on the image by pressing and holding the button. This area defines the region to be cropped.

Manual mode¶

Alternatively, enable the Manual mode to enter specific cropping coordinates directly in the dialog.

The select area button takes those coordinates from the image: draw a rectangle exactly as in the Interactive mode and, instead of cropping, its coordinates are placed into the edit boxes. The dialog switches to the Manual mode, so that the area can be adjusted before cropping.

ROI-based cropping¶

You can also crop based on a selected region of interest by enabling the from ROI mode, which uses ROIs defined in the ROI panel.

Start cropping¶

To perform the crop, click the Crop button to crop the current dataset.
Alternatively, use the Crop to button to copy the cropped dataset to another buffer. Buffers are managed via the buttons at the top of the Datasets panel.

Note

To restore a cropped dataset into the original, use the Fuse into existing mode of Dataset Chunking → Stitch dataset available at Ribbon → Home → Dataset Chunking → Stitch dataset.

BigData datasets¶

When the active dataset is a BigData (OME-Zarr v3 pyramidal) dataset, additional controls become available in the Settings panel of the Crop dialog.

Only the requested region is read, so cropping a small area of a very large or remote dataset costs no more than the area itself.

Note

Cropping a BigData or Virtual dataset cannot be undone. Crop into another container from the Destination dropdown to keep the original.

Zarr pyramid level¶

The Zarr pyramid level dropdown selects which resolution level of the pyramid is used as the source for the crop region. Level s0 is the full-resolution data; s1, s2, … are progressively downsampled versions (e.g. s1 is typically 2× downsampled in XY).

The label below the dropdown shows the corresponding voxel downsampling factors (XYZ) for the selected level, e.g. Zarr downsampling scales (XYZ): 2 x 2 x 1 for s1.

Tip

Choosing a coarser pyramid level (s1, s2, …) produces a smaller in-memory dataset but with proportionally larger voxel sizes.

Output type¶

The Output type dropdown controls what format the cropped data is saved in. It is enabled only when the source dataset is BigData.

Standard¶

The cropped region is loaded from the selected pyramid level into memory as a Standard (in-memory) dataset. The resulting dataset resides entirely in RAM and can be worked on with all standard MIB tools.

The voxel size of the output reflects the selected pyramid level - cropping at s1 doubles the voxel size relative to s0.

BigData¶

The cropped region is written to a new OME-Zarr v3 pyramid on disk. A file-save dialog appears to choose the destination .zarr3 folder.

Warning

A confirmation dialog is shown before the file picker because this operation writes a new pyramid to disk and replaces the dataset in the destination buffer.

Output files created:

File	Contents
`<name>.zarr3`	OME-Zarr v3 image pyramid (always written)
`Labels_<name>.zarr3`	Packed segmentation model pyramid - only created when a BigData model exists in the source dataset

The model pyramid (Labels_<name>.zarr3) is written at full resolution (s0) and all coarser levels are derived automatically. Material names, colors, and material count are copied from the source model.

After writing, the cropped BigData dataset is loaded into the destination buffer automatically. If a model was written, it is attached to the buffer immediately - no manual model loading is required.

The output zarr3 file is highlighted in the Directory Contents panel after the operation completes.

Note

The source BigData file on disk is not modified or deleted. Only the destination buffer is updated.

Back to MIB | User interface | Ribbon | Dataset