We propose a hybrid super-voxel segmentation pipeline that combines soft memberships from fuzzy
c-means (FCM) with spatial regularization via graph cuts on a region adjacency graph (RAG). Our hybrid
achieves 0.70 mean IoU at 42 fps (synthetic), outperforming SLIC-only (0.55) and FCM-only (0.52)
under the same budget. The result is spatially coherent clusters that respect object boundaries while
preserving soft assignment information at real-time performance.
@article{Achanta2012SLIC,
title={SLIC Superpixels Compared to State-of-the-art Superpixel Methods},
author={Achanta, Radhakrishna and Shaji, Appu and Smith, Kevin and Lucchi, Aur{\'e}lien and Fua, Pascal and S{\"u}sstrunk, Sabine},
journal={IEEE TPAMI},
year={2012}, volume={34}, number={11}, pages={2274--2282}
}
@book{Bezdek1981FCM,
title = {Pattern Recognition with Fuzzy Objective Function Algorithms},
author = {Bezdek, James C.},
year = {1981}, publisher = {Springer}
}
@inproceedings{Boykov2001ICCV,
title={Interactive Graph Cuts for Optimal Boundary \& Region Segmentation of Objects in N-D Images},
author={Boykov, Yuri and Jolly, Marie-Pierre},
booktitle={ICCV}, year={2001}, pages={105--112}
}
@article{Boykov2001PAMI,
title={Fast Approximate Energy Minimization via Graph Cuts},
author={Boykov, Yuri and Veksler, Olga and Zabih, Ramin},
journal={IEEE TPAMI}, year={2001}, volume={23}, number={11}, pages={1222--1239}
}
@article{ShiMalik2000PAMI,
title={Normalized Cuts and Image Segmentation},
author={Shi, Jianbo and Malik, Jitendra},
journal={IEEE TPAMI}, year={2000}, volume={22}, number={8}, pages={888--905}
}
@article{Felzenszwalb2004IJCV,
title={Efficient Graph-Based Image Segmentation},
author={Felzenszwalb, Pedro F. and Huttenlocher, Daniel P.},
journal={IJCV}, year={2004}, volume={59}, number={2}, pages={167--181}
}
@article{skimage2014,
title={scikit-image: image processing in {P}ython},
author={van der Walt, St{\'e}fan and Sch{\"o}nberger, Johannes L. and Nunez-Iglesias, Juan and Boulogne, Fran{\c{c}}ois and Warner, Joshua D. and others},
journal={PeerJ}, year={2014}, volume={2}, pages={e453}
}