katalytic-images

Don't use in production yet. I will probably introduce backwards incompatible changes

Process images with less boilerplate and more flexibility than you've ever dreamed of.

• loading images in a specific colorspace (RGB, HSV, grayscale, and more) with a single command
• converting image colorspaces
• The functions accept numpy arrays and Pillow images as input and return the same type
• defining the shapes to be drawn on top of an image in a declarative way as a list of dicts and passing it to draw()
• Many more (TODO: Link to tocs)

Example (TODO)

Installation

First, install opencv, which is not added as dependency so you can install the contrib/headless versions

if [[ ! $(pip freeze | grep -Pi 'opencv') ]]; then
    pip install opencv-python
fi

Then, install this package

pip install katalytic-images

Roadmap

• make pillow an optional dependency.
  • setup_load_image() should pick opencv if pillow is not available
• image thresholding and masking operations
• interactive data exploration widgets (maybe as part of another package)
• higher level API on top of opencv
• utilities for video processing

Contributing

We appreciate any form of contribution, including but not limited to:

• Code contributions: Enhance our repository with your code and tests.
• Feature suggestions: Your ideas can shape the future development of our package.
• Architectural improvements: Help us optimize our system's design and API.
• Bug fixes: Improve user experience by reporting or resolving issues.
• Documentation: Help us maintain clear and up-to-date instructions for users.