Header logo is

Partitioning of Image Datasets using Discriminative Context Information

2008

Conference Paper

ei


We propose a new method to partition an unlabeled dataset, called Discriminative Context Partitioning (DCP). It is motivated by the idea of splitting the dataset based only on how well the resulting parts can be separated from a context class of disjoint data points. This is in contrast to typical clustering techniques like K-means that are based on a generative model by implicitly or explicitly searching for modes in the distribution of samples. The discriminative criterion in DCP avoids the problems that density based methods have when the a priori assumption of multimodality is violated, when the number of samples becomes small in relation to the dimensionality of the feature space, or if the cluster sizes are strongly unbalanced. We formulate DCP‘s separation property as a large-margin criterion, and show how the resulting optimization problem can be solved efficiently. Experiments on the MNIST and USPS datasets of handwritten digits and on a subset of the Caltech256 dataset show that, given a suitable context, DCP can achieve good results even in situation where density-based clustering techniques fail.

Author(s): Lampert, CH.
Book Title: CVPR 2008
Journal: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2008)
Pages: 1-8
Year: 2008
Month: June
Day: 0
Publisher: IEEE Computer Society

Department(s): Empirical Inference
Bibtex Type: Conference Paper (inproceedings)

DOI: 10.1109/CVPR.2008.4587448
Event Name: IEEE Computer Society Conference on Computer Vision and Pattern Recognition
Event Place: Anchorage, AK, USA

Address: Los Alamitos, CA, USA
Digital: 0
Language: en
Organization: Max-Planck-Gesellschaft
School: Biologische Kybernetik

Links: PDF
Web

BibTex

@inproceedings{5083,
  title = {Partitioning of Image Datasets using Discriminative Context Information},
  author = {Lampert, CH.},
  journal = {Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2008)},
  booktitle = {CVPR 2008},
  pages = {1-8},
  publisher = {IEEE Computer Society},
  organization = {Max-Planck-Gesellschaft},
  school = {Biologische Kybernetik},
  address = {Los Alamitos, CA, USA},
  month = jun,
  year = {2008},
  doi = {10.1109/CVPR.2008.4587448},
  month_numeric = {6}
}