Back to my home page.

On Region Merging: the Statistical Soundness of Fast Sorting, with Applications

Frank Nielsen and Richard Nock

Abstract:
This work explores a statistical basis for a process often described in computer vision: image segmentation by region merging following a particular order in the choice of regions. We exhibit a particular blend of algorithmics and statistics whose error is, as we formally show, close to the best possible. This approach can be approximated in a very fast segmentation algorithm for processing images described using most common numerical feature spaces. Simple modifications of the algorithm allow to cope with occlusions and/or hard noise levels. Experiments on grey-level and color images, obtained with a short C-code, display the quality of the segmentation obtained.
Download the PDF paper here (1958 Kb size, 8 pages, 8 figures) © IEEE.
 
The original publication is available here from IEEE Proceedings.
Bibtex entry:
@InProceedings{,
  title = {{On Region Merging: the Statistical Soundness of Fast Sorting, with Applications}},
  author = {Frank Nielsen and Richard Nock}, 
  booktitle = {Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition},
  pages     = {"II:19--26"},
  year      = {2003},
  publisher = {IEEE Computer Society},
  address   = {Los Alamitos, {CA}}
}

Examples of segmented color images:

Software codes:

The software includes options for occlusions handling, setting the number of statistical distributions (parameter Q), add noise to image (salt and pepper, transmission, Gaussian), output layout, etc. 

The software is given ONLY for research purposes and is provided "AS IS" without any guarantee.

Related publications:
 
Frank Nielsen and Richard Nock,
On Region Merging: the Statistical Soundness of Fast Sorting, with Applications,
IEEE International Conference on Computer Vision and Pattern Recognition (CVPR),
Volume 2, pp. 19-26, 2003.
 

Previous work:

© Copyright notice.
Last updated, 2003.