Two Dimensional Filter Design Using Evolutionary Optimization

In the last few years design of two dimensional grown sufficient zest among researchers. The design of two dimensional finite impulse response (FIR) filters can be expound as a non-linear optimization problem. The constraints are high and estimation of large number of parameters is needed, especiall case of two dimensional finite impulse response filters. In order to improve performance we have used Binary Cat Swarm Optimization (BCSO) in which some concepts are introduced to bring down ripples.


INTRODUCTION
Two dimensional(2D) digital filter have found wide applications in different areas like image processing [1], seismic signal processing, nuclear test detection, sonar, radar and radio astrology [2]. In the area image processing, 2D FIR filters are preferred because of their inherently stable nature, no phase distortion and easy realization using fast Fourier Transform. They are used for image contrast enhancement, denoising [3], [4], image deblurring and image restoration [5]. The classical techniques for designing 2D filters are extension of 1D techniques, these include windowing [6], frequency sampling chebysev approximation [8], transformation [2] and least squares. Windowing is the simplest technique for magnitude response but it fails in linear phase of 2D filters [6]. In classical techniques some optimization based approaches has been successfully applied for 2D FIR filter design.
Online @ www.ijtsrd.com | Volume -2 | Issue -5 | Jul-Aug In the last few years design of two dimensional has grown sufficient zest among researchers. The design of two dimensional finite impulse response (FIR) linear optimization problem. The constraints are high and estimation of large number of parameters is needed, especially in case of two dimensional finite impulse response filters. In order to improve performance we have used Binary Cat Swarm Optimization (BCSO) in which some concepts are introduced to bring down ripples.
Finite impulse response, optimization, Two dimensional(2D) digital filter have found wide applications in different areas like image processing , seismic signal processing, nuclear test detection, . In the area of image processing, 2D FIR filters are preferred because of their inherently stable nature, no phase distortion and easy realization using fast Fourier Transform. They are used for image contrast image deblurring and The classical techniques for designing 2D filters are extension of 1D techniques, , frequency sampling [7], McClellan [2] and least squares. Windowing is the simplest technique for magnitude response but it addition to the classical techniques some optimization based approaches has been successfully applied for 2D FIR Optimization based filter design techniques aims at meeting the desired specifications by minimizing a predefined objective function. The techniques reported in this regards can be broadly classified into two groups, i.e., classical gradient and heuristi evolutionary techniques. Gradient based techniques; inspite of having faster convergence, quite often gets trapped into local minima. Also the final solution to these techniques is very sensitive to the initial parameterization. Evolutionary optimization because of their ability to mimic the intelligence of natural selection and adoption process found in biological species, have been successfully applied for the solution of complex nonlinear optimization problem in different engineering fields includ filter design. In this regard, the optimization techniques adopted for 2D FIR filter includes simulated annealing [9], clonal section algorithm genetic algorithm(GA) [11] optimization [12].Most of the above reported techniques suffer from the limitations of tuning , selection of large number of control parameters and large execution time. In this context, a recently developed evolutionary technique i.e., modified binary cat swarm optimization has been applied in the present work for the design of FIR filters.

METHODOLOGY
The Binary cat swarm optimization is an optimization algorithm that imitates the natural behavior of cats. Cats have curiosity about objects in motion and have a great hunting ability. It might be thought that cats spend most of the time resting, but in fact they are constantly alert and moving slowly. This behavior corresponds to the seeking mode. Furthermore, when cats detect a prey, they spend lots of energy bec Optimization based filter design techniques aims at meeting the desired specifications by minimizing a predefined objective function. The techniques reported in this regards can be broadly classified into two groups, i.e., classical gradient and heuristic evolutionary techniques. Gradient based techniques; inspite of having faster convergence, quite often gets trapped into local minima. Also the final solution to these techniques is very sensitive to the initial parameterization. Evolutionary optimization (EA) because of their ability to mimic the intelligence of natural selection and adoption process found in biological species, have been successfully applied for the solution of complex nonlinear optimization problem in different engineering fields including FIR filter design. In this regard, the optimization techniques adopted for 2D FIR filter includes , clonal section algorithm [10], [11], cat swarm .Most of the above reported techniques suffer from the limitations of tuning , selection of large number of control parameters and large execution time. In this context, a recently developed evolutionary technique i.e., modified ation has been applied in the present work for the design of FIR filters.
The Binary cat swarm optimization is an optimization algorithm that imitates the natural behavior of cats. Cats have curiosity about objects in motion and have great hunting ability. It might be thought that cats spend most of the time resting, but in fact they are constantly alert and moving slowly. This behavior corresponds to the seeking mode. Furthermore, when cats detect a prey, they spend lots of energy because