Skip to main content

Visualize 3D numpy arrays in the browser.

Project description

PyPI version

microviewer

Multiplatform 3D numpy image browser based viewer.

from microviewer import view, hyperview, objects

view(numpy_image) # for gray and color images
view(numpy_image, seg=True) # for segmentation labels
view(numpy_image, seg=True, port=8082)

hyperview(image, labels) # interactive overlay

# uses vtk to visualize meshes, bounding boxes, 
# skeletons, and point clouds
objects([ mesh, skeleton, bbox, point_cloud ])
uview image.npy.gz # view as image
uview labels.npy.gz --seg # view as segmentation
uview image.npy.gz labels.npy.gz # view image w/ overlay
uview image.npy.gz --paint # view image w/ blank overlay

Segmentation display in microviewer

Visualize 3D numpy arrays in your browser without difficult installation procedures or reformatting your data. The code is CPU based and the image is uncompressed in memory. You're limited to images that are at most 2^31 bytes large (~2.1 GB) due to browser limitations.

Features

  • 3 axis visualization of 3D images.
  • Grayscale images with segmentation overlay.
  • Segmentation selection with brush tools.
  • Direct voxel painting.
  • Save segmentation as .npy or .ckl, an advanced compresssion format.
  • Undo/Redo

3D Image Support

  • 8-bit grayscale 3D images
  • color images (including 3 channel 3D images)
  • floating point images
  • boolean images
  • segmentation labels
  • .npy, .ckl, .nrrd, or .nii format

For .ckl, .nrrd, and .nii formats, you must install crackle-codec, pynrrd, and nibabel respectively. All are installed by the "all_formats" optional installation argument.

3D Object Support

  • osteoid Skeletons
    • Color By: Connected Component, Radius, Cross Sectional Area
  • zmesh and CloudVolume Meshes
  • Point Clouds
  • Bounding Boxes

Example

import numpy as np
import zmesh
import kimimaro
import crackle
import microviewer
from osteoid.lib import Bbox

labels = crackle.load("connectomics.npy.ckl.gz", label=62347522)
resolution = np.array([16,16,40])

skel = kimimaro.skeletonize(
  labels, 
  teasar_params={
    'scale': 3,
    'pdrf_exponent': 8,
  },
  anisotropy=resolution,
)[1]

mesher = zmesh.Mesher(resolution)
mesher.mesh(labels)
mesh = mesher.get(1)

bbox = Bbox([0,0,0], np.array(labels.shape) * resolution)

microviewer.objects([ bbox, mesh, skel ], skeleton_color_by='radius')

3D object display in microviewer via vtk

Installation

pip install microviewer

By default, only .npy files are supported. Additional formats are supported via optional downloads.

Possible optional arguments:

all_formats, nii, nrrd, ckl, jxl,
objects,
all

Note: jxl and objects have relatively heavy downloads.

pip install "microviewer[all]"

History

This microviewer package has been a part of CloudVolume since 2018, but is now broken out into its own package for more flexible wider use. Microviewer uses a modified version of https://github.com/seung-lab/data-cube-x/ (2016) to represent the array in Javascript, which was originally developed for eyewire.org.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

microviewer-1.17.0.tar.gz (1.5 MB view details)

Uploaded Source

Built Distribution

microviewer-1.17.0-py3-none-any.whl (160.2 kB view details)

Uploaded Python 3

File details

Details for the file microviewer-1.17.0.tar.gz.

File metadata

  • Download URL: microviewer-1.17.0.tar.gz
  • Upload date:
  • Size: 1.5 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.12.3

File hashes

Hashes for microviewer-1.17.0.tar.gz
Algorithm Hash digest
SHA256 84997259a32261c5464181e5e7af5165596bacbf688fa206b0d060512b97ced9
MD5 ab7fb5e7e3db5ecdb9257e937c6f8b66
BLAKE2b-256 fba59e5e6e338113d2426810ba6d2e6e7e44b9d597fb76d45739e1829c106524

See more details on using hashes here.

File details

Details for the file microviewer-1.17.0-py3-none-any.whl.

File metadata

  • Download URL: microviewer-1.17.0-py3-none-any.whl
  • Upload date:
  • Size: 160.2 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.12.3

File hashes

Hashes for microviewer-1.17.0-py3-none-any.whl
Algorithm Hash digest
SHA256 cf2ec271fe04c1acafe49aa002a7600182fdb5cd9d03c7efcaeec07a21b4ed0e
MD5 e5acf4915204c34551f1ee5f22c3ce24
BLAKE2b-256 98219e0758d596ddbb82f5ac926c8c80c992e99b7d4a4cf91124582d38d166a8

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page