Presentation 2005-12-16
Morphological Associative Memory and its unknown Recalling Procedure Derived from Principle of Duality
Sadayuki Murashima, Ke ZHANG,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Morphological associative memory is known as a robust memory which can recall stored pattern from contaminated with noise. In this paper we present a new recalling process which is different from Ritter's original recalling process. In Ritter's morphological associative memory the signals pass through the memory matrix Mzz with kernel Z which includes a few fired bits in order to utilize the property that M is robust for dilative noise. There is an another recalling procedure in which the signals pass through the memory matrix Wrr with kernel R in which almost all pixels are fired bits. We call Ritter's kernel as normal one and what we presented as the reverse kernel. By mean of several numerical experiments the new recalling process is also effective and turned out to be founded on the same basis of morphological associative memory as Ritter's. These two recalling procedures are both effective and do not denote any relative merits.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) morphology / associative memory / recalling procedure / principle of duality
Paper # NLP2005-87
Date of Issue

Conference Information
Committee NLP
Conference Date 2005/12/9(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Nonlinear Problems (NLP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Morphological Associative Memory and its unknown Recalling Procedure Derived from Principle of Duality
Sub Title (in English)
Keyword(1) morphology
Keyword(2) associative memory
Keyword(3) recalling procedure
Keyword(4) principle of duality
1st Author's Name Sadayuki Murashima
1st Author's Affiliation Faculty of Engineering, Kagoshima University()
2nd Author's Name Ke ZHANG
2nd Author's Affiliation Ricoh Software Inc.
Date 2005-12-16
Paper # NLP2005-87
Volume (vol) vol.105
Number (no) 483
Page pp.pp.-
#Pages 6
Date of Issue