Normalized Cuts and Image Segmentation
Abstract
We propose a novel approach for solving the perceptual grouping problem in vision. Rather than focusing on local features and their consistencies in the image data, our approach aims at extracting the global impression of an image. We treat image segmentation as a graph partitioning problem and propose a novel global criterion, the normalized cut, for segmenting the graph. The normalized cut criterion measures both the total dissimilarity between the different groups as well as the total similarity within the groups. We show that an efficient computational technique based on a generalized eigenvalue problem can be used to optimize this criterion. We have applied this approach to segmenting static images and found results very encouraging.
Cite
Text
Shi and Malik. "Normalized Cuts and Image Segmentation." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1997. doi:10.1109/CVPR.1997.609407Markdown
[Shi and Malik. "Normalized Cuts and Image Segmentation." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1997.](https://mlanthology.org/cvpr/1997/shi1997cvpr-normalized/) doi:10.1109/CVPR.1997.609407BibTeX
@inproceedings{shi1997cvpr-normalized,
title = {{Normalized Cuts and Image Segmentation}},
author = {Shi, Jianbo and Malik, Jitendra},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year = {1997},
pages = {731-737},
doi = {10.1109/CVPR.1997.609407},
url = {https://mlanthology.org/cvpr/1997/shi1997cvpr-normalized/}
}