| | |
Summary: Non-Parametric Probabilistic Image Segmentation
Marco Andreetto Lihi Zelnik-Manor Pietro Perona
Department of Electrical Engineering
California Institute of Technology
Pasadena, CA, 91125, USA
{marco, lihi, perona}@caltech.edu
Abstract
We propose a simple probabilistic generative model for
image segmentation. Like other probabilistic algorithms
(such as EM on a Mixture of Gaussians) the proposed model
is principled, provides both hard and probabilistic cluster
assignments, as well as the ability to naturally incorporate
prior knowledge. While previous probabilistic approaches
are restricted to parametric models of clusters (e.g., Gaus-
sians) we eliminate this limitation. The suggested approach
does not make heavy assumptions on the shape of the clus-
ters and can thus handle complex structures. Our experi-
ments show that the suggested approach outperforms previ-
ous work on a variety of image segmentation tasks.
1. Introduction
|
