xarray-based tools for image/video processing


Keywords
image, video, labelling, patch, extraction, data, labelling-images
License
Other
Install
pip install ximage==0.0.4

Documentation

đŸ“Ļ Welcome to ximage

Deployment PyPI Conda
Activity PyPI Downloads Conda Downloads
Python Versions Python Versions
Supported Systems Linux macOS Windows
Project Status Project Status
Build Status Tests Lint Docs
Linting Black Ruff Codespell
Code Coverage Coveralls Codecov
Code Quality Codefactor Codebeat Codacy CodeScene
License License
Community Slack GitHub Discussions
Citation DOI

Slack | Docs

ℹī¸ Software Overview

The software currently enables to:

  • label n-dimensional xarray objects
  • extract patches around n-dimensional labels
  • extract patches from n-dimensional xarray objects

Join the Slack Workspace to meet the community !

🚀 Quick Start

ximage provides an easy-to-use interface to manipulate image, videos and n-dimensional arrays with classical image processing techniques.

The ximage xarray accessor provides a convenient way to labelling and extract patches in n-dimensional arrays !

Image labelling
min_value_threshold = 1
max_value_threshold = np.inf
min_area_threshold = 5
max_area_threshold = np.inf
footprint = None
sort_by = "area"
sort_decreasing = True
label_name = "label"

### Label xarray object
xr_obj = da.ximage.label(
    min_value_threshold=min_value_threshold,
    max_value_threshold=max_value_threshold,
    min_area_threshold=min_area_threshold,
    max_area_threshold=max_area_threshold,
    footprint=footprint,
    sort_by=sort_by,
    sort_decreasing=sort_decreasing,
    label_name=label_name,
)
Extract patches around labels
# Output Options
n_patches = 10
n_labels = None
labels_id = None
highlight_label_id = False
# Patch Extraction Options
patch_size = (100, 100)
centered_on = "label_bbox"
padding = 0
n_patches_per_label = np.Inf
n_patches_per_partition = 1
# Tiling/Sliding Options
partitioning_method = None # "tiling" / "sliding"
n_partitions_per_label = None
kernel_size = None
buffer = 0
stride = None
include_last = True
ensure_slice_size = True
debug = True
verbose = True

da_patch_gen = xr_obj.ximage.label_patches(
    label_name=label_name,
    patch_size=patch_size,
    variable=variable,
    # Output Options
    n_patches=n_patches,
    n_labels=n_labels,
    labels_id=labels_id,
    highlight_label_id=highlight_label_id,
    # Patch Extraction Options
    centered_on=centered_on,
    padding=padding,
    n_patches_per_label=n_patches_per_label,
    n_patches_per_partition=n_patches_per_partition,
    # Tiling/Sliding Options
    partitioning_method=partitioning_method,
    n_partitions_per_label=n_partitions_per_label,
    kernel_size=kernel_size,
    buffer=buffer,
    stride=stride,
    include_last=include_last,
    ensure_slice_size=ensure_slice_size,
    debug=debug,
    verbose=verbose,
)

📖 Explore the ximage documentation

To discover all ximage utilities, please read the software documentation.

🛠ī¸ Installation

conda

ximage can be installed via conda on Linux, Mac, and Windows. Install the package by typing the following command in the terminal:

conda install ximage

In case conda-forge is not set up for your system yet, see the easy to follow instructions on conda-forge.

pip

ximage can be installed also via pip on Linux, Mac, and Windows. On Windows you can install WinPython to get Python and pip running.

Install the ximage package by typing the following command in the terminal:

pip install ximage

To install the latest development version via pip, see the documentation.

💭 Feedback and Contributing Guidelines

If you aim to contribute or discuss the future development of ximage, we highly suggest to join the Slack Workspace

Feel free to also open a GitHub Issue or a GitHub Discussion specific to your questions or ideas.

✍ī¸ Contributors

Citation

You can cite the ximage software by:

Ghiggi Gionata & Son Pham-Ba . ghiggi/ximage. Zenodo. https://doi.org/10.5281/zenodo.8131552

If you want to cite a specific version, have a look at the Zenodo site.

License

The content of this repository is released under the terms of the MIT license.

📚 Requirements: