vesskel 3.6.0

2D and 3D Skeletonization and Graph-based Analysis

VesSkel

Vessel Skeletonization and Graph-Based Phenotype Analysis in Retinal Fundus Images

Installation

uv sync                  # core only
uv sync --extra dev      # + test tools
uv sync --extra napari   # + napari GUI
uv sync --all-extras     # everything

Napari

uv sync --extra napari && uv run napari

Open a manual1 TIFF from the HRF folder, then run Lee94 Thinning from the VesSkel plugin menu to see the skeleton.

Inside the Analyze Vessels widget, tune extraction settings and use Save Config to export a reusable JSON preset.

CLI

Use the same JSON preset exported from napari to batch-process images.

vesskel config-init --out config.json
vesskel validate-config --config config.json
vesskel run --input HRF/manual1 --config config.json --out outputs

CLI outputs:

• outputs/summary.csv with one feature row per image
• Optional per-image skeleton outputs (default: .npy)
• Optional per-image branch tables when output.write_branch_csv=true

Shell completions

# zsh
eval "$(vesskel completions zsh)"

# bash
eval "$(vesskel completions bash)"

# PowerShell
vesskel completions powershell | Out-String | Invoke-Expression

Add the appropriate line to your shell rc for persistent tab-completion.

Tests

uv sync --extra dev && uv run pytest                     # all tests
uv sync --extra dev && uv run pytest -m "not slow"       # skip regression tests

• 2D regression - thinning + feature extraction on all 45 HRF samples, compared against saved baselines
• 3D regression - thinning + features on a brain volume (from scikit-image), same baseline approach
• 3D comparison - vesskel lee94_thin vs skimage.morphology.skeletonize on the brain volume, asserting identical output

First run (or --update-baseline) generates baselines in tests/skeletons/ and tests/features/.

Dataset

This project uses the High-Resolution Fundus (HRF) Image Database, established by a collaborative research group to support comparative studies on automatic segmentation algorithms on retinal fundus images.

The database contains 45 images total:

• 15 images of healthy patients
• 15 images of patients with diabetic retinopathy
• 15 images of glaucomatous patients

Binary gold standard vessel segmentation images and field of view (FOV) masks are available for each image.

License

Budai, Attila; Bock, Rüdiger; Maier, Andreas; Hornegger, Joachim; Michelson, Georg. Robust Vessel Segmentation in Fundus Images. International Journal of Biomedical Imaging, vol. 2013, 2013

The HRF dataset is released under the Creative Commons 4.0 Attribution License.

For more information, visit the HRF Image Database.