SplitTouchingObjects

SplitTouchingObjects applies distance-transform watershed inside each non-zero label mask and returns a sequentially relabeled output image.

Inputs are labels and min_distance. Outputs are output_labels and object_count.

Minimal Example

from bioimageflow_core import Arguments
from bioimageflow_segmentation_tools import SplitTouchingObjects

result = SplitTouchingObjects().process_row(
    Arguments(labels="clumped.tif", output_labels="split.tif", min_distance=5)
)

Inputs

• labels: 2D label image.

• min_distance: minimum marker spacing in pixels.

Outputs

• output_labels: relabeled split objects.

• object_count: number of output labels.

Dependencies and Core Libraries

imageio, NumPy, SciPy distance transforms, and scikit-image watershed helpers.

Assumptions

Each input label is a clump where distance peaks can approximate object centers.

Expected Results

Synthetic clumped disks split into separate sequential output labels.

Failure Modes

Unreadable labels, invalid marker settings, unsupported dimensions, and write failures raise errors.