Skip to content
/ SVCD Public

A Python implementation of the Spatially Varying Color Distributions for Interactive Multi-Label Segmentation

License

Notifications You must be signed in to change notification settings

drgHannah/SVCD

Folders and files

NameName
Last commit message
Last commit date

Latest commit

0798c16 · Feb 10, 2023

History

3 Commits
Apr 14, 2022
Apr 14, 2022
Apr 14, 2022
Apr 14, 2022
Apr 14, 2022
Feb 10, 2023
Apr 14, 2022

Repository files navigation

Spatially Varying Color Distributions for Interactive Multi-Label Segmentation

An unofficial Python implementation of the paper Spatially Varying Color Distributions for Interactive Multi-Label Segmentation by Claudia Nieuwenhuis and Daniel Cremers. This implementation was done in collaboration with @unlikelymaths.

Get Started

  • Install
    Please install this repository by:

      pip install ./SVCD/
    
  • Data The required data consists of an RGB image and a suitable image with user-drawn scribbles. In the image with scribbles, black values and pixels with alpha<1 are assumed to be unknown, besides each individual color is treated as a class.

Usage

Please also have a look into samples.

from svcd.svcd import SVCDutils, SVCDSegmentation

image_path = ...
scribble_path = ...
save_path = ...

# load image and scribbles
image = SVCDutils.load_image(image_path)
scribble, colors = SVCDutils.load_segmentation_from_image(scribble_path)

# run algorithm
svcd = SVCDSegmentation(max_iter=1000)
segmentation, energies = svcd(image, scribble)  

# save segmentation
SVCDutils.save_segmentation_as_image(save_path, segmentation, colors)

About

A Python implementation of the Spatially Varying Color Distributions for Interactive Multi-Label Segmentation

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages