This repository offers the code for the Adaptels algorithm with a Python interface. One can produce superpixels for grey, color, as well as images with any other number of channels. A demo file is provided, which should be easy to use. When using for the first time, the C files need to be compiled using:
python compile_adaptels_lib.py
The demo can then be run using:
python AdaptelsDemo.py
The output of the demo should be as follows:
The Adaptels algorithm is different from conventional superpixel algorithms in that it generates compact segments of varying size that adapt to the local image scale. In regions of low information it generates large superpixels and vice versa. It simple to use since it takes just one parameter (which can be set to 60 for sRGB images). It is also computationally very efficient.
If you use the code, please cite the following publication:
"Scale-Adaptive Superpixels", R. Achanta and P Marquez-Neila and P. Fua and S. Süsstrunk, Color and Imaging Conference (CIC26), 2018
@article{adaptels_2018,
title = {Scale-Adaptive Superpixels},
author = {Achanta, Radhakrishna and Marquez Neila, Pablo and Fua, Pascal and Süsstrunk, Sabine},
publisher = {Society for Imaging Science and Technology},
journal = {26th Color and Imaging Conference Final Program and Proceedings},
pages = {1-6(6)},
year = {2018},
}