README.md

metadata

license: cc-by-nc-4.0
tags:
  - vesuvius-challenge
  - herculaneum
  - ct
  - lasagna
  - ome-zarr
pretty_name: PHerc0175A lasagna field (villa predict3d)

PHerc0175A lasagna field

villa's lasagna predict3d output for the whole PHerc0175A scan, in villa's own layout, so you can point lasagna or the spiral fit at it without running the model yourself.

See the same height in 3D, with the center line and the normals, in the Umbilici Space (its sheet at z 6,400; this preview is at z 6,372).

Get it

hf download SharpMeow/pherc0175a-lasagna --repo-type dataset --local-dir pherc0175a-lasagna
cd pherc0175a-lasagna && sha256sum -c SHA256SUMS
for f in parts/*.tar; do tar -xf "$f"; done
tar -xf PHerc0175A-lasagna-pyramid-fix.tar   # last: joins the level-4 pyramid across tiles

You end up with PHerc0175A.lasagna.json and PHerc0175A_{cos,grad_mag,nx,ny}.ome.zarr.

Load it

import numpy as np
import zarr

cos = zarr.open("PHerc0175A_cos.ome.zarr/1", mode="r")  # every 2nd scan voxel
nx = zarr.open("PHerc0175A_nx.ome.zarr/2", mode="r")    # every 4th scan voxel
ny = zarr.open("PHerc0175A_ny.ome.zarr/2", mode="r")

z, y, x = 1536, 6000, 4608                             # full-resolution scan voxels
patch = cos[z // 2, y // 2 - 256:y // 2 + 256, x // 2 - 256:x // 2 + 256]
a = (float(nx[z // 4, y // 4, x // 4]) - 128) / 127    # normals are stored as uint8, v * 127 + 128
b = (float(ny[z // 4, y // 4, x // 4]) - 128) / 127
c = np.sqrt(max(0.0, 1 - a * a - b * b))               # nz is not stored

load_example.py is a runnable version of this snippet; it was run on the real parts before upload.

Stream it, no download

The same field, unpacked, is in the public bucket SharpMeow/lasagna-fields under PHerc0175A/representations/predictions/lasagna/20250521115057-lasagna-de45f011/, so any zarr reader can open it by URL:

import zarr   # zarr 3, with fsspec and aiohttp

base = "https://huggingface.co/buckets/SharpMeow/lasagna-fields/resolve/PHerc0175A/representations/predictions/lasagna/20250521115057-lasagna-de45f011"
cos = zarr.open_array(base + "/PHerc0175A_cos.ome.zarr/1", mode="r", zarr_format=2)

Region downloads, a Neuroglancer link and the methods for all ten fields are on the Lasagna Fields page.

What it is

• Scan: PHerc0175A, 20250521115057-8.640um-1.2m-116keV-masked.zarr, level 0, Vesuvius Challenge open data
• Model: scrollprize/lasagna model_current.pt at revision de45f011 (sha256 4cc0c40a...)
• Code: ScrollPrize/villa lasagna at db3dbea2, lasagna-preprocess predict3d --tile-size 192, default chunks and levels
• Channels: cos at OME-Zarr level 1 (2x down); grad_mag, nx, ny at level 2 (4x down); each with its pyramid
• Run: 64 parts on Hugging Face Jobs (RTX PRO 6000). Every part edge sits on a chunk line for levels 1 to 3. Level 4 chunks that straddle a tile edge were rebuilt from the joined level 3 with villa's own omezarr_pyramid, and a seam check found no step at any tile edge (worst ratio 1.26 against a 1.5 limit). In grad_mag the rebuilt level 4 differs from what the parts wrote in 15,978 of 57.7 million voxels (0.03%), every one a 0 against a small value at the edge of the masked region; nx and ny differ in one voxel each. Finer levels are exactly as the parts wrote them.

Notes

Raw model output; nobody has checked it by eye. Regions the masked scan leaves empty are absent (fill 0). The model ran on the scan at its native 8.64 um with no resampling; we have not tested whether another input scale works better.

Related

A center line (umbilicus) for PHerc0175A, and for every other scan on the prize page, is in SharpMeow/vesuvius-center-lines.

The Lasagna tab of the Umbilici Space shows a small sample of this field in your browser: cos at nine heights and the surface normals as needles, with the center line through them. Its X-ray tab streams the PHerc0175A CT itself.

Credit and terms

The scan is Vesuvius Challenge open data, CC BY-NC 4.0; this derived field is shared under the same terms. Model and code are ScrollPrize/villa's lasagna, credit to its authors. Shared by Chase Hendrick.