Let H be a fixed directed graph on h vertices, let G be a directed graph on n vertices and suppose that at least n2 edges have to be deleted from it to make it H-free. We show tha...
"Pattern recognition techniques are concerned with the theory and algorithms of putting abstract objects, e.g., measurements made on physical objects, into categories. Typical...
Many semi-supervised learning algorithms only
deal with binary classification. Their extension to the
multi-class problem is usually obtained by repeatedly
solving a set of bina...
Many computer vision problems can be formulated in
a Bayesian framework with Markov Random Field (MRF)
or Conditional Random Field (CRF) priors. Usually, the
model assumes that ...
We propose an unconventional but highly effective approach
to robust fitting of multiple structures by using statistical
learning concepts. We design a novel Mercer kernel
for t...