Application of wavelet-matched filters to holographic associative memories

We propose a novel method for improving association abilities of a holographic associative memory by using wavelet-matched filtering. Based on an edge-enhancement and muitiscaling properties of the wavelet transform, a sharp correlation peak between a partial input and a stored datum can be obtained...

Full description

Bibliographic Details
Main Authors: Widjaja, Joewono, Tomita, Yasuo, Abdul Rahman, Abdul Wahab
Format: Article
Language:English
Published: Elsevier 1996
Subjects:
Online Access:http://irep.iium.edu.my/38324/
http://irep.iium.edu.my/38324/
http://irep.iium.edu.my/38324/
http://irep.iium.edu.my/38324/1/Application_of_wavelet-matched_filters_to_holographic_associative_memories.pdf
Description
Summary:We propose a novel method for improving association abilities of a holographic associative memory by using wavelet-matched filtering. Based on an edge-enhancement and muitiscaling properties of the wavelet transform, a sharp correlation peak between a partial input and a stored datum can be obtained. This leads to higher association abilities. Autoand hetero-association operations are demonstrated experimentally.