A Graphical Model for Simultaneous Partitioning and Labeling
Abstract
In this work we develop a graphical model for describing probability distributions over labeled partitions of an undirected graph which are conditioned on observed data. We show how to efficiently perform exact inference in these models, by exploiting the structure of the graph and adapting the sum-product and max-product algorithms. We demonstrate our approach on the task of segmenting and labeling hand-drawn ink fragments, and show that a significant performance increase is obtained by labeling and partitioning simultaneously. 1
Cite
Text
Cowans and Szummer. "A Graphical Model for Simultaneous Partitioning and Labeling." Proceedings of the Tenth International Workshop on Artificial Intelligence and Statistics, 2005.Markdown
[Cowans and Szummer. "A Graphical Model for Simultaneous Partitioning and Labeling." Proceedings of the Tenth International Workshop on Artificial Intelligence and Statistics, 2005.](https://mlanthology.org/aistats/2005/cowans2005aistats-graphical/)BibTeX
@inproceedings{cowans2005aistats-graphical,
title = {{A Graphical Model for Simultaneous Partitioning and Labeling}},
author = {Cowans, Philip J. and Szummer, Martin},
booktitle = {Proceedings of the Tenth International Workshop on Artificial Intelligence and Statistics},
year = {2005},
pages = {73-80},
volume = {R5},
url = {https://mlanthology.org/aistats/2005/cowans2005aistats-graphical/}
}