Clifford networks: novel neural architectures for multidimensional data

Clifford algebra based neural networks offer an alternative model to the more traditional architectures allowing: concise abstract modelling of complex multidimensional problem domains, reduced network size compared with traditional network configurations, and improved training algorithm efficiency...

Full description

Saved in:
Bibliographic Details
Published in6th International Conference on Image Processing and its Applications pp. 670 - 675
Main Authors Howells, G, Bisset, D.L, Fairhurst, M.C
Format Conference Proceeding
LanguageEnglish
Published London IEE 1997
Subjects
Algebra
Applied sciences
Artificial intelligence
Computer science; control theory; systems
Computer vision and image processing techniques
Electric, optical and optoelectronic circuits
Electronics
Exact sciences and technology
Neural computing techniques
Neural nets
Neural networks
Optical information, image and video signal processing
Pattern recognition
Pattern recognition. Digital image processing. Computational geometry
multidimensional data
training algorithm efficiency
algebra
colour image processing
complex multidimensional problem domains
network size
network configuration
Clifford networks
backpropagation
Clifford algebra based neural networks
neural net architecture
image colour analysis
neural architectures
abstract modelling
Online AccessGet full text

Cover

More Information
Summary:Clifford algebra based neural networks offer an alternative model to the more traditional architectures allowing: concise abstract modelling of complex multidimensional problem domains, reduced network size compared with traditional network configurations, and improved training algorithm efficiency compared with conventional training algorithms. Colour image processing is used as an example.
ISBN:085296692X
9780852966921
ISSN:0537-9989
DOI:10.1049/cp:19970979