A new approach to avoiding the local extrema trap

McCabe, A., and Trevathan, J. (2007) A new approach to avoiding the local extrema trap. ANZIAM Journal, 48. pp. 474-493.

[img]
Preview
PDF
Download (186Kb)
View at Publisher Website: http://anziamj.austms.org.au/ojs/index.p...

Abstract

The Extremum Consistency algorithm avoids local maxima and minima in a specialised domain. The most notable difference between this approach and others is that it places a greater importance on the width or consistency of an extremum than on its height or depth (amplitude). Short term, high amplitude extrema are encountered in many typical situations (such as noisy environments or due to hardware inaccuracies) and cause problems with system accuracy. The Extremum Consistency algorithm is far less susceptible to these situations than hill climbing, convolution, thresholding, and tends to produce higher quality results. We describes the algorithm and present results from practical experimentation, which illustrates its superiority over other forms of local extrema avoidance in three real world applications.

Item ID: 2258
Item Type: Article (Refereed Research - C1)
Keywords: handwritten signature verification; convolution; hill climbing
ISSN: 1446-8735
Date Deposited: 29 May 2009 04:56
FoR Codes: 08 INFORMATION AND COMPUTING SCIENCES > 0804 Data Format > 080499 Data Format not elsewhere classified @ 100%
SEO Codes: 89 INFORMATION AND COMMUNICATION SERVICES > 8999 Other Information and Communication Services > 899999 Information and Communication Services not elsewhere classified @ 100%
Downloads: Total: 36
Last 12 Months: 2
More Statistics

Actions (Repository Staff Only)

Item Control Page Item Control Page