RoiMaskValidator

class zrad.preprocessing.masks.RoiMaskValidator(aggregation_dimension=None)

Validate ROI mask geometry for 2D, 2.5D, or 3D feature extraction.

The validator enforces minimum ROI extent and voxel-count requirements before feature extraction. In slice-wise modes, invalid slices are removed; in 3D mode, the full volume must satisfy the checks.

Parameters:

aggregation_dimension ({"2D", "2.5D", "3D", None}, optional) – Feature aggregation mode that determines the geometric validity checks. "3D" validates the full ROI volume. Other non-None values are treated as slice-wise modes and remove invalid slices. If None, the mask is copied without geometric validation.

Methods

apply(mask)	Validate a mask and return a validated mask copy.
get_params()	Return validation parameters mapped to their configured values.

RoiMaskValidator.apply(mask)

Validate a mask and return a validated mask copy.

Parameters:

mask (Image) – Binary ROI mask to validate.

Returns:

validated_mask – Copy of the input mask. In slice-wise modes, slices that fail the minimum requirements are set to zero.

Return type:

Image

RoiMaskValidator.get_params()

Return validation parameters mapped to their configured values.

Returns:

params – Dictionary containing aggregation_dimension.

Return type:

dict