Pattern Recognition Using Neural and Functional Networks free download online
Title: Pattern Recognition Using Neural and Functional Networks Author(s): Vasantha Kalyani David Pages: 196 Publisher: Springer Publication date: 2008 Language: English Format: PDF ISBN-10: 3540851291 ISBN-13: Description: Preface
Biologically inspired computing is different from conventional computing. It has
a different feel; often the terminology does not sound like its talking about
machines. The activities of this computing sound more human than mechanistic
as people speak of machines that behave, react, self-organize, learn, generalize,
remember and even to forget. Much of this technology tries to mimic natures
approach in order to mimic some of natures capabilities. They have a rigorous,
mathematical basis and neural networks for example have a statistically valid
set on which the network is trained.
Two outlines are suggested as the possible tracks for pattern recognition. They
are neural networks and functional networks. Neural Networks (many interconnected
elements operating in parallel) carry out tasks that are not only beyond
the scope of conventional processing but also cannot be understood in the same
terms. Imaging applications for neural networks seem to be a natural fit. Neural
networks love to do pattern recognition. A new approach to pattern recognition
using microARTMAP together with wavelet transforms in the context of hand
written characters, gestures and signatures have been dealt. The Kohonen Network,
Back Propagation Networks and Competitive Hopfield Neural Network
have been considered for various applications.
Functional networks, being a generalized form of Neural Networks where functions
are learned rather than weights is compared with Multiple Regression Analysis
for some applications and the results are seen to be coincident.
New kinds of intelligence can be added to machines, and we will have the
possibility of learning more about learning. Thus our imaginations and options
are being stretched. These new machines will be fault-tolerant, intelligent and
self-programming thus trying to make the machines smarter. So as to make those
who use the techniques even smarter.
Chapter 1 is a brief introduction to Neural and Functional networks in the
context of Pattern recognition using these disciplines Chapter 2 gives a review
of the architectures relevant to the investigation and the development of these
technologies in the past few decades.
VIII Preface
Chapter 3 begins with the look at the recognition of handwritten alphabets
using the algorithm for ordered list of boundary pixels as well as the Kohonen
Self-Organizing Map (SOM). Chapter 4 describes the architecture of the
MicroARTMAP and its capability.
Chapter 5 the MicroARTMAP is augmented with a moment based feature
extractor and applied to character recognition in this chapter. Chapter 6 illustrates
the use of wavelet transforms together with MicroARTMAP for character
recognition. The microARTMAP gave solutions to problems in civil engineering
like classification of soil problem, finding the load from yield pattern of a plate
and finding earthquake parameters from a given response spectrum.
Chapter 7 MicroARTMAP and Back Propagation Network have been compared
and presented for gesture recognition and signature verification. Solving
scheduling problem by means of a Competitive Hopfield Neural Network are discussed
in Chapter 8 Multiple Regression methods considered as a recognizer is
compared with functional networks in solving certain problems as shown in
Chapter 9. Conclusion and further applications are suggested in Chapter 10.
# Hardcover: 196pages
# Publisher: Springer; 1 edition (December 1, 2008)
# Language: English
# ISBN-10: 3540851291
# ISBN-13: 978-3540851295
Pattern Recognition Using Neural and Functional Networks free download links:
Title: Pattern Recognition Using Neural and Functional Networks