forement 0.0.1

A collection of unsupervised clustering tools for foreground segmentation

FOREground segMENTation

A collection of unsupervised clustering tools for foreground segmentation of an image to isolate an object of interest.

Requirements:

• ultralytics
• opencv-python
• orchard-bouman

Developed in Python 3.10, but anything over Python 3.7 should work

Installation:

pip install forement

Usage:

To use from the command line use the command:

python forement.py -image_path -method -k -channel -output_dir -save_image -show_image

Arguments:

• image_path, type=str, help=Input the path to the image to perform clustering on.
• method, type=str, help=Method to segment images by, options include "sam", "kmeans", "ob", "em", "ob-em".
• k, type=int | None, default=None, help=How many clusters to generate. For method "ob" or "ob-em" clusters is 2^k.
• channel, type=int, default=0, help=Color channel to use when running mode counts on images.
• output_dir, type=str | None, default=None, help=Folder to save images to. If None, created images are saved in the folder of the original image.
• save_image, type=bool, default=False, help=Whether or not to save the segmented image.
• show_image, type=bool, default=False, help=Whether or not to display the segmented image.

Available methods:

• kmeans: "kmeans"
• Orchard Bouman: "ob"
• Expectation Maximization: "em"
• Expectation Maximization using Orchard Bouman: "ob-em"
• SAM (Segment anything model): "sam"

In addition to the command line call, all functions can be imported to your own scripts and used as part of a larger program.