Titlebook: Information Theory in Computer Vision and Pattern Recognition; Francisco Escolano,Pablo Suau,Boyán Bonev Textbook 2009 Springer-Verlag Lon

arboretum

滋養(yǎng)

極力證明

resuscitation

photophobia

猜忌

Registration, Matching, and Recognition,e instantiated to intensity distributions. Therefore, image registration can be posed as finding the (constrained) transformation that holds the maximal dependency between distributions. This rationale opens the door to the quest for new measures rooted in mutual information.

鳴叫

APEX

Feature Selection and Transformation,undant, and some could introduce noise, or be irrelevant. In some problems the number of features is very high and their dimensionality has to be reduced in order to make the problem tractable. In other problems feature selection provides new knowledge about the data classes. For example, in gene se

不如樂(lè)死去

Classifier Design, building unique but deep trees, in favor of a bunch of shallow trees. This is the keypoint of the chapter, the emergence of ., complex classifiers built in the aggregation/combination of simpler ones, and the role of IT in their design. In this regard, the method adapted to images is particularly i

Generalize

978-1-4471-5693-2Springer-Verlag London Ltd., part of Springer Nature 2009