Integrating Local Classifiers through Nonlinear Dynamics on Label Graphs with an Application to Image Segmentation
We present a new method to combine possibly inconsistent locally (piecewise) trained conditional models p(y|x) into pseudo-samples from a global model. Our method does not require training of a CRF, but instead generates samples by iterating forward a weakly chaotic dynamical system. The new method is illustrated on image segmentation tasks where classifiers based on local appearance cues are combined with pairwise boundary cues.
Download: pdf
Text Reference
Yutian Chen, Andrew Gelfand, Charless Fowlkes, and Max Welling. Integrating local classifiers through nonlinear dynamics on label graphs with an application to image segmentation. In Proc. of the International Confernece on Computer Vision. 2011.BibTeX Reference
@inproceedings{ChenGFW_ICCV_2011,author = "Chen, Yutian and Gelfand, Andrew and Fowlkes, Charless and Welling, Max",
title = "Integrating Local Classifiers through Nonlinear Dynamics on Label Graphs with an Application to Image Segmentation",
booktitle = "Proc. of the International Confernece on Computer Vision",
year = "2011",
tag = "grouping"
}