Segmentation Cleanup Module

Correct and measure a labeled segmentation stack in one integrated editor. Paint fixes by hand, run automated cleanup over the whole volume, quantify the result, and save it as a new file.

The Segmentation Cleanup module works on a two-step flow: select a segmentation, then edit and quantify it. All operations run against a server-side working copy. The original file is never modified; every change is written out as a new, tracked file.

Key Features

• Manual painting with brush, eraser, and flood-fill tools

• Automated cleanup: merge or remove classes, fill holes, remove small components, smooth boundaries

• Per-class quantification: voxel counts, physical volume, connected components, surface area

• Report export to CSV, registered in the file browser

• Non-destructive: edits apply to a working copy, results are saved as new files

Supported Input

• Segmentation results from U-Net inference

• Annotation masks uploaded or created in the Annotation Tool

• Any labeled TIFF stack with at least one non-zero class

Outputs are tagged as segmentation results, so the Mesh Generation and Stitching pickers accept them directly.

Workflow

1. Select: choose a segmentation or annotation mask to edit and measure

2. Edit & Quantify: paint corrections, apply automated cleanup, and review per-class metrics

3. Create report: write the quantification to CSV files in the workspace

4. Save as New File: write the cleaned stack as a tracked output, then continue editing or open it in the Image Viewer