Rutuparna Panda

Tsallis entropy based optimal multilevel thresholding using cuckoo search algorithm

Abstract In this paper, optimal thresholds for multi-level thresholding in an image are obtained by maximizing the Tsallis entropy using cuckoo search algorithm. The method is considered as a constrained optimization problem. The solution is obtained through the convergence of a meta-heuristic search algorithm. The proposed algorithm is tested on standard set of images. The results are then compared with that of bacteria foraging optimization (BFO), artificial bee colony (ABC) algorithm, particle swarm optimization (PSO) and genetic algorithm (GA). Results are analyzed both qualitatively and quantitatively. It is observed that our results are also encouraging in terms of CPU time and objective function values.

Fractional generalized splines and signal processing

This paper presents the concept of fractional generalized splines, which is an extension of the idea of Unsers fractional splines. The first part of this paper describes a method for construction of fractional generalized splines through evaluating fractional finite differences. The main key to our approach is to provide an additional tuning parameter α by using a generating function, which is the solution of the Laguerres nth-order differential equation. The second part of the paper deals with characterization of these functions in both time and frequency domain and shows how to use these results for construction of wavelet bases in L2 for signal processing applications. This paper also present simulation results to reveal the suitability of the proposed basis functions for signal approximation.

Edge magnitude based multilevel thresholding using Cuckoo search technique

Abstract Multilevel thresholding technique is popular and extensively used in the field of image processing. In this paper, a multilevel threshold selection is proposed based on edge magnitude of an image. The gray level co-occurrence matrix (second order statistics) of the image is used for obtaining multilevel thresholds by optimizing the edge magnitude using Cuckoo search technique. New theoretical formulation for objective functions is introduced. Key to our success is to exploit the correlation among gray levels in an image for improved thresholding performance. Apart from qualitative improvements the method also provides us optimal threshold values. Results are compared with histogram (first order statistics) based between-class variance method for multilevel thresholding. It is observed that the results of our proposed method are encouraging both qualitatively and quantitatively.

Face recognition using bacterial foraging strategy

Abstract This article presents an efficient algorithm for LDA-based face recognition with the selection of optimal principal components using E-coli Bacterial Foraging Optimization Technique. Different methods were suggested in the literature to select the largest eigenvalues and their corresponding eigenvectors for linear discriminant analysis (LDA). Some researchers have suggested eliminating the three largest eigenvalues to avoid the effect under varying illumination conditions. But, there is no unified approach for selecting optimal eigenvalues to enhance the performance of an algorithm. In this context, a GA–PCA algorithm has been proposed to select optimal eigenvalues and their corresponding eigenvectors in LDA. They proposed a fitness function to find the optimal eigenvectors using the Genetic Algorithm (GA). However, the crossover method used results in differences in offspring, and mutation never allowed them for a physical dispersal of the child in a chosen area. This prevents us in selecting optimal eigenvectors for improvising accuracy of the face recognition algorithm. This has motivated the authors to develop a new algorithm called BFO-Fisher which uses a nutrient concentration function (cost function) for optimization. In this work, the cost function is maximized through hill climbing via a type of biased random walk which is not possible in GA. Here the proposed BFO-Fisher algorithm offers us two distinct additional advantages—(i) the proposed algorithm can supplement the features of GA, and (ii) the random bias incorporated into the BFO algorithm guides us to move in the direction of increasingly favorable environment, which is desirable. In this experiment, both Yale and UMIST Databases are used for the performance evaluation. Experimental results presented in this article reveal the fact that about 3% (Rank 1) improvement can be achieved as compared to the GA-Fisher algorithm.

A study on fuzzy clustering for magnetic resonance brain image segmentation using soft computing approaches

A novel hybrid GA-BFO-FCM algorithm for MR brain image segmentation is presented.Optimal boundary point detection (OBPD) method using GA-BFO is investigated.Inbuilt constraints and repair mechanism is incorporated in our method.Reformulation of the objective function is well suited for optimization.Improved segmentation results are obtained. Our method is robust for noisy images. This paper presents a novel idea of intracranial segmentation of magnetic resonance (MR) brain image using pixel intensity values by optimum boundary point detection (OBPD) method. The newly proposed (OBPD) method consists of three steps. Firstly, the brain only portion is extracted from the whole MR brain image. The brain only portion mainly contains three regions-gray matter (GM), white matter (WM) and cerebrospinal fluid (CSF). We need two boundary points to divide the brain pixels into three regions on the basis of their intensity. Secondly, the optimum boundary points are obtained using the newly proposed hybrid GA-BFO algorithm to compute final cluster centres of FCM method. For a comparison, other soft computing techniques GA, PSO and BFO are also used. Finally, FCM algorithm is executed only once to obtain the membership matrix. The brain image is then segmented using this final membership matrix. The key to our success is that we have proposed a technique where the final cluster centres for FCM are obtained using OBPD method. In addition, reformulated objective function for optimization is used. Initial values of boundary points are constrained to be in a range determined from the brain dataset. The boundary points violating imposed constraints are repaired. This method is validated by using simulated T1-weighted MR brain images from IBSR database with manual segmentation results. Further, we have used MR brain images from the Brainweb database with additional noise levels to validate the robustness of our proposed method. It is observed that our proposed method significantly improves segmentation results as compared to other methods.

Generalized B-spline signal processing

Abstract This paper describes an efficient generalized B-spline filtering technique for the processing and representation of signals in terms of continuous generalized B-spline basis functions. Different types of generalized B-spline of any degree are introduced in a more generalized framework. An efficient generalized B-spline filtering technique for band-limited signal is presented. It is shown that the entire filtering process is equivalent to lowpass filtering of the band-limited signal. In the first stage of the filtering process, we consider the problem of extracting spline coefficients with down-sampling with a factor of m . In the second stage, we consider the reconstruction of the original signal through up-sampling with an optional zooming factor m . Normalized gain versus frequency response characteristics for some of the generalized B-spline filters are plotted for interpretation of the quality of generalized B-spline filtering. A computer simulation result is presented to show the antialiasing property of generalized B-spline filtering for band-limited signals.

An efficient algorithm for gray level image enhancement using cuckoo search

This paper presents an efficient algorithm for gray level image enhancement using Cuckoo search (CS). The results are compared with Particle Swarm Optimization (PSO). The basic idea is to treat image enhancement as an optimization problem and then solve it using CS. It is observed that the proposed method provides better results than existing techniques.

Design of 1-D and 2-D recursive filters using crossover bacterial foraging and Cuckoo search techniques

Recently, there has been an increasing interest on the application of the evolutionary algorithms in order to solve the drawbacks of traditional filter design methods. Unlike classical methods, they offer the advantage of not requiring a good initial estimate of filter parameters to proceed. This paper presents design of one-dimensional (1-D) and two-dimensional (2-D) recursive filters using crossover bacterial foraging (COBFO) and Cuckoo Search (CS) techniques. Design of 1-D and 2-D recursive filters is considered here as a constrained optimization problem to ensure stability. The solution is obtained through convergence of a biased random search using crossover bacterial foraging optimization technique to ensure quality. A faster solution is also obtained through the convergence of a meta heuristic search technique called the Cuckoo search technique. Inbuilt constraint handling capability makes our proposal attractive in the design of recursive filters. Results are compared with genetic algorithm (GA) and bacteria foraging optimization (BFO) techniques.

A novel adaptive crossover bacterial foraging optimization algorithm for linear discriminant analysis based face recognition

Our evolutionary face recognition algorithm provides improved recognition rate.A novel adaptive crossover bacterial foraging optimization algorithm is proposed.The new algorithm improves the objective function values.Optimal dimension reduction is achieved using ACBFO-Fisher algorithm.ACBFO-Fisher algorithm search optimal eigenvectors to improve accuracy. This paper presents a modified bacterial foraging optimization algorithm called adaptive crossover bacterial foraging optimization algorithm (ACBFOA), which incorporates adaptive chemotaxis and also inherits the crossover mechanism of genetic algorithm. First part of the research work aims at improvising evaluation of the optimal objective function values. The idea of using adaptive chemotaxis is to make it computationally efficient and crossover technique is to search nearby locations by offspring bacteria. Four different benchmark functions are considered for performance evaluation. The purpose of this research work is also to investigate a face recognition algorithm with improved recognition rate. In this connection, we propose a new algorithm called ACBFO-Fisher. The proposed ACBFOA is used for finding optimal principal components for dimension reduction in linear discriminant analysis (LDA) based face recognition. Three well-known face databases, FERET, YALE and UMIST, are considered for validation. A comparison with the results of earlier methods is presented to reveal the effectiveness of the proposed ACBFO-Fisher algorithm.

Least squares generalized B-spline signal and image processing

Abstract This paper presents generalized B-spline interpolation and approximation techniques. The first part of this paper describes a technique for the continuous representation of discrete signals in terms of generalized B-splines (direct B-spline transform) and for interpolative signal reconstruction (indirect B-spline transform) with an expansion factor m . The key to our approach is to use a family of generalized B-spline filters. It has been shown that the scale factor can be modified to tune the approximated filter bandwidths. The second part of this paper makes use of these results to describe an equivalent filtering interpretation (low-pass filtering followed by a generalized B-spline interpolator) of generalized B-splines together with least squares generalized B-spline signal approximation methods. This part also concern with new theoretical results showing factorization of the transfer functions of the higher degree ( n >3) generalized B-spline filters. Experimentally, it is seen that data compression with reduced amount of error can be achieved using these least squares generalized B-spline filtering techniques.