Pattern Recognition using Neural and Functional Networks

Sundaramoorthy Rajasekaran, Vasantha Kalyani David

ISBN 10: 3540851291 / ISBN 13: 9783540851295

Published by Springer-Verlag GmbH, 2008

About this Item

Zustand: Sehr gut | Seiten: 184 | Sprache: Englisch | Produktart: Bücher.

Seller Inventory # 4562959/12

Bibliographic Details

• Title: Pattern Recognition using Neural and Functional Networks
• Publisher: Springer-Verlag GmbH
• Publication Date: 2008
• Language: English
• ISBN 10: 3540851291
• ISBN 13: 9783540851295
• Binding: Hardcover
• Condition: Sehr gut
• Seller catalogs:
  • Bücher

About this title

Synopsis

Pattern Recognition Using Neural and Functional Networks Pattern Recognition Using Neural and Functional Networks

From the Back Cover

The concept of pattern is universal in intelligence and discovery. The patterns in biological data contain knowledge. Discrimination of signal pattern allows personal identification by voice, hand writing, finger prints, facial images, recognition of speech, written characters and also scenes in images like identification of military targets based on radar, infrared, and video images. Possibilities are enormous in geologic, climatic, meteorologic, personality, cultural, historical, spectral, electromagnetic as well as from microscopic images of cells to macroscopic images of regions of the earth obtained from satellite scans and radio telescope images of galaxies. It is up to the researcher in some area to glean the essentials and begin to explore the classification and recognition of patterns in data that will lead to discoveries of associations and cause – effect relationships. Two outlines are suggested as the possible tracks for pattern recognition. They are neural networks and functional networks. A new approach to pattern recognition using microARTMAP and 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.