Many approaches to object recognition are founded on probability theory, and can be broadly characterized as either generative or discriminative according to whether or not the di...
In this paper we present a fully automated approach to the segmentation of pediatric brain tumors in multi-spectral 3-D magnetic resonance images. It is a top-down segmentation app...
Michael Wels, Gustavo Carneiro, Alexander Aplas,...
A Bayesian framework for deformable pattern classification has been proposed in [1] with promising results for isolated handwritten character recognition. Its performance, however...
The fact that objects in the world appear in different ways depending on the scale of observation has important implications if one aims at describing them. It shows that the not...
As the amount of textual information grows explosively in various kinds of business systems, it becomes more and more desirable to analyze both structured data records and unstruc...