Rehab F. Abdel-Kader

Genetically Improved PSO Algorithm for Efficient Data Clustering

Clustering is an important research topic in data mining that appears in a wide range of unsupervised classification applications. Partitional clustering algorithms such as the k-means algorithm are the most popular for clustering large datasets. The major problem with the k-means algorithm is that it is sensitive to the selection of the initial partitions and it may converge to local optima. In this paper, we present a hybrid two-phase GAI-PSO+k-means data clustering algorithm that performs fast data clustering and can avoid premature convergence to local optima. In the first phase we utilize the new genetically improved particle swarm optimization algorithm (GAI-PSO) which is a population-based heuristic search technique modeled on the hybrid of cultural and social rules derived from the analysis of the swarm intelligence (PSO) and the concepts of natural selection and evolution (GA). The GAI-PSO combines the standard velocity and position update rules of PSOs with the ideas of selection, mutation and crossover from GAs. The GAI-PSO algorithm searches the solution space to find the optimal initial cluster centroids for the next phase. The second phase is a local refining stage utilizing the k-means algorithm which can efficiently converge to the optimal solution. The proposed algorithm combines the ability of the globalized searching of the evolutionary algorithms and the fast convergence of the k-means algorithm and can avoid the drawback of both. The performance of the proposed algorithm is evaluated through several benchmark datasets. The experimental results show that the proposed algorithm is highly forceful and outperforms the previous approaches such as SA, ACO, PSO and k-means for the partitional clustering problem.

An efficient eye detection and tracking system based on particle swarm optimization and adaptive block-matching search algorithm

Abstract The problem of eye detection and tracking in video sequences is very important for a large number of applications ranging from face recognition to gaze tracking. Eye detection and tracking are challenging due to a variety of factors such as eye-blinking, partially closed eyes, and oblique face orientations which tend to significantly limit the efficiency of most eye trackers. In this paper, an efficient eye detection and tracking system is presented to overcome these limitations. The proposed system switches between the particle swarm optimization (PSO) based deformable multiple template matching algorithm and the adaptive block-matching search algorithm to improve the efficiency and robustness of the tracking system. For eye detection, PSO-based deformable multiple template matching is employed to estimate the best candidate of the center of the eyes within an image of the video sequence with the highest accuracy. For eye tracking the block-matching algorithm with adaptive search area is utilized to reduce the computational time required to perform the PSO-based algorithm. Experimental results on the standard VidTIMIT database show that the proposed method outperforms the deformable template matching based methods such as genetic and PSO. Moreover, it achieves better performance compared to model-based methods such as the statistical active appearance model (AAM) method and the edge projections based method in terms of accuracy and computational complexity.

Particle swarm optimization for constrained instruction scheduling

Instruction scheduling is an optimization phase aimed at balancing the performance-cost tradeoffs of the design of digital systems. In this paper, a formal framework is tailored in particular to find an optimal solution to the resource-constrained instruction scheduling problem in high-level synthesis. The scheduling problem is formulated as a discrete optimization problem and an efficient population-based search technique; particle swarm optimization (PSO) is incorporated for efficient pruning of the solution space. As PSO has proven to be successful in many applications in continuous optimization problems, the main contribution of this paper is to propose a new hybrid algorithm that combines PSO with the traditional list scheduling algorithm to solve the discrete problem of instruction scheduling. The performance of the proposed algorithms is evaluated on a set of HLS benchmarks, and the experimental results demonstrate that the proposed algorithm outperforms other scheduling metaheuristics and is a promising alternative for obtaining near optimal solutions to NP-complete scheduling problem instances.

3D Face Compression and Recognition using Spherical Wavelet Parametrization

In this research an innovative fully automated 3D face compression and recognition system is presented. Several novelties are introduced to make the system performance robust and efficient. These novelties include: First, an automatic pose correction and normalization process by using curvature analysis for nose tip detection and iterative closest point (ICP) image registration. Second, the use of spherical based wavelet coefficients for efficient representation of the 3D face. The spherical wavelet transformation is used to decompose the face image into multi-resolution sub images characterizing the underlying functions in a local fashion in both spacial and frequency domains. Two representation features based on spherical wavelet parameterization of the face image were proposed for the 3D face compression and recognition. Principle component analysis (PCA) is used to project to a low resolution sub-band. To evaluate the performance of the proposed approach, experiments were performed on the GAVAB face database. Experimental results show that the spherical wavelet coefficients yield excellent compression capabilities with minimal set of features. Haar wavelet coefficients extracted from the face geometry image was found to generate good recognition results that outperform other methods working on the GAVAB database.

Fuzzy Particle Swarm Optimization with Simulated Annealing and Neighborhood Information Communication for Solving TSP

In this paper, an effective hybrid algorithm based on Particle Swarm Optimization (PSO) is proposed for solving the Traveling Salesman Problem (TSP), which is a well-known NP- complete problem. The hybrid algorithm combines the high global search efficiency of fuzzy PSO with the powerful ability to avoid being trapped in local minimum. In the fuzzy PSO system, fuzzy matrices were used to represent the position and velocity of the particles in PSO and the operators in the original PSO position and velocity formulas were redefined. Two strategies were employed in the hybrid algorithm to strengthen the diversity of the particles and to speed up the convergence process. The first strategy is based on Neighborhood Information Communication (NIC) among the particles where a particle absorbs better historical experience of the neighboring particles. This strategy does not depend on the individual experience of the particles only, but also the neighbor sharing information of the current state. The second strategy is the use of Simulated Annealing (SA) which randomizes the search algorithm in a way that allows occasional alterations that worsen the solution in an attempt to increase the probability of escaping local optima. SA is used to slow down the degeneration of the PSO swarm and increase the swarms diversity. In SA, a new solution in the neighborhood of the original one is generated by using a designed λ search method. A new solution with fitness worse than the original solution is accepted with a probability that gradually decreases at the late stages of the search process. The hybrid algorithm is examined using a set of benchmark problems from the TSPLIB with various sizes and levels of hardness. Comparative experiments were made between the proposed algorithm and regular fuzzy PSO, SA, and basic ACO. The computational results demonstrate the effectiveness of the proposed algorithm for TSP in terms of the obtained solution quality and convergence speed.

Orientation of multiple principal axes shapes using efficient averaging method

In this paper we propose an effective, low computational cost technique to find the orientation of shapes that have several non-equally separated axes of symmetry. In our technique we define a simple method to calculate the average angle of the shapes axes of symmetry. The axes of symmetry of the shape could be detected using any of the well known techniques reported in the literature. In the proposed technique we use the edge points of the shape to have the ability to deal with natural pictures like coins. The internal edges are used in addition to the external boundary edges to increase the orientation detection capabilities of the algorithm. First, the edge map of the image is extracted by applying Canny edge detector [15]. Second, the center of the object is detected by calculating the average of the vertical and horizontal coordinates of the points of the edge map. Third, the total perpendicular absolute distances from the edge map points to the line that passes through the center point with specified angle are calculated. These calculations are repeated with different angles to find the angles of the minimum peaks of the calculated distances. Finally, if the shape has more than one minimum peak we use our averaging method to get the dominant direction angle of the shape or the shape orientation. By using this technique we only use the first moment of inertia and do not have to use any higher orders to reduce the computational cost.

A hybrid rotation-invariant face recognition system using Log-Polar Transform

The recognition of human faces, especially those with different orientations is a challenging and important problem in image analysis and classification. This paper proposes an effective schema for rotation invariant face recognition using Log-Polar Transform and Discrete Cosine Transform combined features. The rotation invariant feature extraction for a given face image involves applying the log-polar transform to eliminate the rotation effect and to produce a row shifted log-polar image. The discrete cosine transform is then applied to eliminate the row shift effect and to generate the low-dimensional feature vector. A particle swarm optimization based feature selection algorithm is utilized to search the feature vector space for the optimal feature subset. Evolution is driven by a fitness function defined in terms of maximizing the between-class separation (scatter index). Experimental results based on the ORL face database using testing data sets for face images with different orientations show that the proposed system outperforms other face recognition methods. The overall recognition rate for the rotated test images being 97%, demonstrating that the extracted feature vector is an effective rotation invariant feature set with minimal set of selected features.

A Hybrid Color Image Quantization Algorithm Based on k-Means and Harmony Search Algorithms

ABSTRACT Color quantization is one of the most important preprocessing stages in many applications in computer graphics and image processing. In this article, a new algorithm for color image quantization based on the harmony search (HS) algorithm is proposed. The proposed algorithm utilizes the clustering method, which is one of the most extensively applied methods to the color quantization problem. Two variants of the algorithm are examined. The first is based on a standalone HS algorithm, and the second is a hybrid algorithm of k-means (KM) and HS. The objective of the hybrid algorithm is to strengthen the local search process and balance the quantization quality and computational complexity. In the first stage, the high-resolution color space is initially condensed to a lower-dimensional color space by multilevel thresholding. In the second stage, the compressed colors are clustered to a palette using the hybrid KMHS to obtain final quantization results. The algorithm aims to design a postclustering quantization scheme at the color-space level instead of the pixel level. This significantly reduces the computational complexity while maintaining the quantization quality. Experimental results on some of the most commonly used test images in the quantization literature demonstrate that the proposed method is a powerful method, suggesting a higher degree of precision and robustness compared to existing algorithms.

Particle swarm optimization for human face recognition

Feature selection (FS) is a global optimization problem in machine learning that reduces the number of features, removes irrelevant, noisy and redundant data, and results in acceptable recognition accuracy. It is the most important step that affects the performance of a pattern recognition system. This paper presents a novel feature selection algorithm based on particle swarm optimization (PSO). The algorithm is applied to coefficients extracted by two feature extraction techniques: the discrete cosine transform (DCT) and the discrete wavelet transform (DWT). The proposed PSO-based feature selection algorithm is utilized to search the feature space for the optimal feature subset where features are carefully selected according to a well defined discrimination criterion. Evolution is driven by a fitness function defined in terms of maximizing the class separation (scatter index). The classifier performance and the length of selected feature vector are considered for performance evaluation using the ORL face database. Experimental results show that the PSO-based feature selection algorithm was found to generate excellent recognition results with the minimal set of selected features.

An improved PSO algorithm with genetic and neighborhood-based diversity operators for the job shop scheduling problem

ABSTRACT The job shop scheduling problem (JSSP) is an important NP-hard practical scheduling problem that has various applications in the fields of optimization and production engineering. In this paper an effective scheduling method based on particle swarm optimization (PSO) for the minimum makespan problem of the JSSP is proposed. New variants of the standard PSO operators are introduced to adapt the velocity and position update rules to the discrete solution space of the JSSP. The proposed algorithm is improved by incorporating two neighborhood-based operators to improve population diversity and to avoid early convergence to local optima. First, the diversity enhancement operator tends to improve the population diversity by relocating neighboring particles to avoid premature clustering and to achieve broader exploration of the solution space. This is achieved by enforcing a circular neighboring area around each particle if the population diversity falls beneath the adaptable diversity threshold. The adaptive threshold is utilized to regulate the population diversity throughout the different stages of the search process. Second, the local search operator based on critical path analysis is used to perform local exploitation in the neighboring area of the best particles. Variants of the genetic well-known operators “selection” and “crossover” are incorporated to evolve stagnated particles in the swarm. The proposed method is evaluated using a collection of 123 well-studied benchmarks. Experimental results validate the effectiveness of the proposed method in producing excellent solutions that are robust and competitive to recent state-of-the-art heuristic-based algorithms reported in literature for nearly all of the tested instances.