VesselBoost - Blood Vessel Segmentation

Vessel segmentation and review in the browser

Process brain vessel maps locally in your browser.

VesselBoost guides NIfTI or DICOM angiography data through optional preprocessing, 3D UNet vessel segmentation, brain masking, interactive review, and reportable local outputs.

Patient images, masks, voxel values, screenshots, and generated outputs stay on this computer.

Local-first workflow

Files are read into browser memory for viewing and analysis.
Model, runtime, and preprocessing assets may be downloaded when a workflow needs them.
Patient-derived files and generated results are not uploaded by the app.
Background execution is possible if this site is added under "Always keep these sites active" in your browser settings.

Downsample (Optional)

Loading volume info...

Downsample Factor i

Step 2: Denoising (Optional)

Denoise the volume to improve segmentation quality. iBilateral is fast with good edge preservation. NLM variants are slower but may produce better results. Processing time depends on volume size.

Method i

Step 3: Vessel Segmentation

Model i

Sliding Window Overlap i

Probability Threshold i

Min Component Size i

Step 4: Brain Extraction

Mask segmentation to brain region. Removes false positives outside the brain.

Method i

Process brain vessel maps locally in your browser.

Local-first workflow

About VesselBoost

Version

Privacy

What this means

How it works

Third-party services

Citations

Deep Learning Framework

Browser Inference

DICOM Conversion

Brain Extraction

Visualization