SNIC superpixels

Simple Non-Iterative Clustering (SNIC) is an improved version of the Simple Linear Iterative Clustering (SLIC) algorithm. SNIC is non-iterative, enforces connectivity from the start, requires less memory, is faster, and yet is a simpler algorithm. On segmentation benchmarks SNIC performs better than the state-of-the-art, including SLIC.

C++ Code with Matlab interface

Please click here to find the C++ code.

Examples

SNIC superpixels

 

SNIC polygons

Reference

If you use SNIC in your work, please cite the following:

 @inproceedings{snic_cvpr17, author = {Achanta, Radhakrishna and Susstrunk, Sabine}, title = {Superpixels and Polygons using Simple Non-Iterative Clustering}, booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)}, year = {2017} }