Abstract: The paper focuses on the problem of the point
correspondence matching in stereo images. The proposed matching
algorithm is based on the combination of simpler methods such as
normalized sum of squared differences (NSSD) and a more complex
phase correlation based approach, by considering the noise and other
factors, as well. The speed of NSSD and the preciseness of the
phase correlation together yield an efficient approach to find the best
candidate point with sub-pixel accuracy in stereo image pairs. The
task of the NSSD in this case is to approach the candidate pixel
roughly. Afterwards the location of the candidate is refined by an
enhanced phase correlation based method which in contrast to the
NSSD has to run only once for each selected pixel.
Abstract: The spread of Web 2.0 has caused user-generated content explosion. Users can tag resources to describe and organize
them. Tag clouds provide rough impression of relative importance of
each tag within overall cloud in order to facilitate browsing among
numerous tags and resources. The goal of our paper is to enrich
visualization of tag clouds. A font distribution algorithm has been
proposed to calculate a novel metric based on frequency and content,
and to classify among classes from this metric based on power
law distribution and percentages. The suggested algorithm has been
validated and verified on the tag cloud of a real-world thesis portal.