ASIAN JOURNAL OF PHYSICS

An International Peer Reviewed Research Journal
Frequency : Monthly,
ISSN : 0971 – 3093
Editor-In-Chief (Hon.) :
Dr. V.K. Rastogi
e-mail:[email protected]
[email protected]

Asian
Journal of Physics

Vol  11, No. 3, July-September, 2002

Proceedings of The National Conference on Lasers & Spectroscopy

CONTENTS

 Vol 11, No 3(2002) 313-344

An investigation of the non-linear properties of correlation flter synthesis and neural netwok design

I Kypraios, R C D Young, P M Birch and C R Chatwin

Industrial Informatics and Manufacturing Systems Centre

School of Engineering  and Information Technology,Brighton BN1 9QT, United Kingdom

Since the introduction of SDF correlation filters vast amounts of work has been done on improving filters’ synthesis. Several  different variations of classic SDF filter have been suggested. The objective is to create filters invariant to distortions from rotation, scaling, translation and inclusion of noise. Recently, the use of wavelets as a pre-processing stage of the images has been explored. Additionally, the application of non-linear operations as part of the pre-processing stages seems promising. This paper investigates those non-linear operations in filters like NL-DOG SDF filter. We also discuss the newly introduced SDF-MACH filter in this paper. Advances in artificial neural network architectures emerged their wide availability in pattern recognition  tasks. We have created a new non-linear neural model able to input analogue patterns. This paper focuses in exploring in detail how any non-linear operations are utilized in our model comparatively to correlate filters. We mathematically prove the non-linear superposition characteristics of the neural netwok. Finally simulation results are given to demonstrate both, the inside synthesis of SDF type correlation filters and the neural network design under the scope of those non-linear operations. We do not intend to provide an analytical comparison of correlation filters and neural networks rather to focus on the understanding of their underlying working functions. ©Anita Publications. All rights reserved.

 Vol 11, No 3(2002) 345-366

Optical Comparator based on FLC over silicon technology

T D Wilkinson and N J New

Cambridge University, Engineering Department

Trumpington St, Cambridge CB2  1PZ,U K

Recent developments have shown that it is possible to create an optical correlator, or comparator, based on ferroelectric liquid crystal spatial light modulator technology. In this paper we present the expansion of this idea to include an in-house FLC over silicon SLM that allows a compact recognizer to be built. The basic architecture is the binary 1/f correlator built around a high-speed 320 ×240 pixel FLC over silicon SLM. Such an architecture can be built with low tolerance opto-mechanics to make a powerful yet robust optical camparator.Experimental and simulation results from this  comparator are presented in this paper alongwith a discussion of how this architecture can be used in more abstract recognition tasks such as motion estimation instead of the more traditional ‘tank-in-the-bush’ scenarios. ©Anita Publications. All rights reserved.

 Vol 11, No 3(2002) 367-389

Similarity suppression algorithm for designing pattern Discrimination filters

David R Selviah and Epaminondas Stamos   

Department of Electronic and Electrical Engineering

University College London, Torrington Place, London, WC1E 7JE, U K

The similarity suppression (SS) algorithm is introduced which designs inner product correlator filters for optical pattern recognition and discrimination. The algorithm which mutually orthogonalises two sets of patterns is compared to the matrix design technique for Linear    Combination Filters (LCF) and is shown to converge to a similar set of filters. Computer simulations for discrimination of a set of 32, 16 ×16 binary, bipolar patterns demonstrate that the designed filters can tolerate 7 dB more noise than matched filters for the same performance and the dynamic range required is reduced by 25 dB making them suitable for use in optical correlators. Moreover, filters designed after only 2 iteration were found to have much better discrimination ability than LCFs in the important case of high noise and low SNR. Their outer products were also found to be lower. © Anita Publications. All rights reserved.