https://github.com/digital-porous-media/dpm_srg

Multi-Seeded Region Growing Segmentation

https://github.com/digital-porous-media/dpm_srg

Science Score: 26.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
  • DOI references
    Found 1 DOI reference(s) in README
  • Academic publication links
  • Academic email domains
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (9.7%) to scientific vocabulary
Last synced: 5 months ago · JSON representation

Repository

Multi-Seeded Region Growing Segmentation

Basic Info
  • Host: GitHub
  • Owner: digital-porous-media
  • License: other
  • Language: C++
  • Default Branch: main
  • Size: 51.8 KB
Statistics
  • Stars: 0
  • Watchers: 1
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created over 1 year ago · Last pushed over 1 year ago
Metadata Files
Readme License

README.md

Seeded Region Growing

Seeded Region Growing (SRG) is a method for image segmentation. It involves selecting initial seed points and expanding these regions by adding neighboring pixels that share similar properties, such as grayscale intensity. The process continues iteratively until the entire image is segmented into distinct labels.

This library is an adaptation of the SRG plugin to Fiji/ImageJ and is based on the algorithm proposed in Adams and Bischof. Our contribution was to translate the original algorithm to C++ and wrap it in Python. We provide this package under the Digital Porous Media organization.

Installation:

dpm_srg is packaged on pypi and can be installed with pip. pip install dpm-srg.

If installing from source, this package requires a C++ compiler.

Usage Example:


This implementation of SRG expects a 2D or 3D grayscale (single color channel) image of type uint8, uint16, or uint32, and a seed image of the same shape as the grayscale image and of type uint8. All seeds should be labeled between 1-255, inclusive as the algorithm fills in points in the seeded image that are labeled 0. The algorithm will return a labeled image of the same shape as the input images of type uint8. Therefore, up to 255 unique labels are possible.

Note that the performance and segmentation accuracy is highly sensitive to the selection of seed points. It is generally better to take a more conservative approach when selecting seed points.

We wrapped each version (2D vs. 3D, dtype) of the template class into individual class instances. The nomenclature is: SRG[2(or 3)]Dunumberofbits (e.g. ```SRG2Du8(),SRG3D_u32()```).

Python Example: ``` import dpm_srg import numpy as np

np.random.seed(130621) image = np.random.randint(0, 256, size=(100, 200, 200), dtype=np.uint16) seeds = np.random.randint(0, 5, size=(100, 200, 200), dtype=np.uint8)

srgobj = dpmsrg.SRG3Du16(image, seeds) srgobj.segment() segmentation = srgobj.getresult() ```

Acknowledgements:

This project includes code concepts adapted from Seeded Region Growing from ijp-toolkit, which is licensed under the LGPL-2.1 License.

Original code can be found here

Owner

  • Name: Digital Porous Media
  • Login: digital-porous-media
  • Kind: organization
  • Email: bcchang@utexas.edu

GitHub Events

Total
Last Year

Packages

  • Total packages: 1
  • Total downloads:
    • pypi 166 last-month
  • Total dependent packages: 0
  • Total dependent repositories: 0
  • Total versions: 3
  • Total maintainers: 1
pypi.org: dpm-srg

Seeded Region Growing Segmentation

  • Versions: 3
  • Dependent Packages: 0
  • Dependent Repositories: 0
  • Downloads: 166 Last month
Rankings
Dependent packages count: 10.2%
Average: 33.9%
Dependent repos count: 57.5%
Maintainers (1)
Last synced: 5 months ago

Dependencies

.github/workflows/build.yml actions
  • actions/checkout v2 composite
  • actions/setup-python v4 composite
pyproject.toml pypi
  • numpy *
setup.py pypi