New Fractal Features for Character Recognition and Image Classification

Linnell, T.A. and Deravi, F. (2005) New Fractal Features for Character Recognition and Image Classification. In: The Institution of Engineering and Technology (IEE). IEE Conference publications. Iee pp. 87-92. ISBN 0863415075 . (The full text of this publication is not available from this repository)

The full text of this publication is not available from this repository. (Contact us about this Publication)
Official URL
http://dx.doi.org/10.1049/cp:20050075

Abstract

In this paper we derive and test some novel fractal based features that are calculated from the fractal compressed representation of an image. We show that these features summarise useful information in the fractal code that can be applied to various types of image classification, including face/non-face classification and character recognition. A preliminary study of two supplementary features that augment the Mapping Vector Accumulator is also presented. Experimental results using these features are provided to indicate their potential for image recognition tasks.

Item Type: Conference or workshop item (Paper)
Additional information: Issue number: CP509
Uncontrolled keywords: fractal features, character recognition, image classification, fractal compressed representation, image representation, fractal code, face classification, nonface classification, mapping vector accumulator, image recognition
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7800 Electronics (see also: telecommunications) > TK7880 Applications of electronics (inc industrial & domestic) > TK7882.P3 Pattern Recognition
Divisions: Faculties > Science Technology and Medical Studies > School of Engineering and Digital Arts > Image and Information Engineering
Depositing User: Yiqing Liang
Date Deposited: 17 Aug 2009 13:53
Last Modified: 14 Jan 2010 14:33
Resource URI: http://kar.kent.ac.uk/id/eprint/8925 (The current URI for this page, for reference purposes)
  • Depositors only (login required):