Radial basis function neural network-based face recognition using firefly algorithm

This paper presents an adaptive technique for obtaining centers of the hidden layer neurons of radial basis function neural network (RBFNN) for face recognition. The proposed technique uses firefly algorithm to obtain natural sub-clusters of training face images formed due to variations in pose, illumination, expression and occlusion, etc. Movement of fireflies in a hyper-dimensional input space is controlled by tuning the parameter gamma (γ) of firefly algorithm which plays an important role in maintaining the trade-off between effective search space exploration, firefly convergence, overall computational time and the recognition accuracy. The proposed technique is novel as it combines the advantages of evolutionary firefly algorithm and RBFNN in adaptive evolution of number and centers of hidden neurons. The strength of the proposed technique lies in its fast convergence, improved face recognition performance, reduced feature selection overhead and algorithm stability. The proposed technique is validated using benchmark face databases, namely ORL, Yale, AR and LFW. The average face recognition accuracies achieved using proposed algorithm for the above face databases outperform some of the existing techniques in face recognition.

     

Development of an algorithm for adaptive compression of indexed images using contextual simulation

An algorithm is proposed for compression of indexed raster images (IRI) based on statistical coding using context simulation. A model and methods are developed in order to create an effective algorithm for compression of indexed graphic information. A detailed consideration is given to methods for increasing the compression ratio. The optimization of the algorithm, including computation parallelization, is presented. The proposed algorithm is compared with other universal and specialized compression algorithms.     

Affine modeling of nonlinear multivariable processes using a new adaptive neural network-based approach

This paper presents a new method for on-line identification of exact affine model for multivariable processes with nonlinear and time-varying behaviors. A self-generating radial basis function (RBF) neural network trained by growing and pruning algorithm for RBF (GAP–RBF) is utilized for deriving the affine model. The extended Kalman filter (EKF) is used for parameter adaptation in the GAP–RBF neural network. The growing and pruning criteria of the original GAP–RBF have been modified with the objective to enhance its performance in on-line identification. Simulation results on two nonlinear multivariable CSTR benchmark problems show an excellent performance of the proposed approach, incorporated with the modified GAP–RBF (MGAP–RBF) neural network, for affine modeling.     

Double EWMA controller using neural network-based tuning algorithm for MIMO non-squared systems

The double exponentially weighted moving average (dEWMA) control method is a popular algorithm for adjusting a process from run to run in semiconductor manufacturing. For MIMO non-squared statistic systems, the singular value decomposition (SVD) method is used for decoupling and the SVD-based dEWMA control scheme is treated as a MIMO extension of dEWMA control design. To enhance the performance and robustness of the linear system in the presence of ramp disturbances and white noises, the neural network-based adaptive algorithm is used to automatically tune the dEWMA controller parameters. Under the specified input patterns, the early stop criterion for the training-validation neural networks, and the stability constraints added in the tuning mechanism, the simulations show that the proposed control technique can effectively improve the means and standard deviations of the process outputs.     

Content adaptive deblocking of artifacts for highly compressed images

Multimedia Tools and Applications - In this paper, a spatial technique for the reduction of blocking artifacts existing within the highly compressed reconstructed images,...     

A neural network-based method for tracking features from satellitesensor images

A new approach for feature tracking on sequential satellite sensor images using neural networks has been developed. The method defines the correspondence problem between features as the minimization of a cost function using a Hopfield neural network. It has been tested on Meteosat radiometer images by tracking a cloud with rotational movement and compared to the maximum cross-correlation method. The Hopfield net was found to be more accurate and faster.     

An RBF neural network-based adaptive control for SISO linearisable nonlinear systems

An RBF neural network-based adaptive control is proposed for Single-Input and Single-Output (SISO) linearisable nonlinear systems in this paper. It is shown that a SISO nonlinear system is first linearised by using the differential geometric approach in the state space, and the linearised nonlinear system is then treated as a partially known system. The known dynamics are used to design a nominal feedback controller to stabilise the nominal system, and an adaptive RBF neural network-based compensator is then designed to compensate for the effects of uncertain dynamics. The main function of the RBF neural network in this work is to adaptively learn the upper bound of the system uncertainty, and the output of the neural network is then used to adaptively adjust the gain of the compensator so that the strong robustness with respect to unknown dynamics can be obtained, and the tracking error between the plant output and the desired reference signal can asymptotically converge to zero. A simulation example is performed in support of the proposed scheme.     

Bounding set calculation for neural network-based output feedback adaptive control systems

Evaluating the bounding set of dynamic systems subject to direct neural-adaptive control is a critical issue in applications where the control system must undergo a rigorous verification process in order to comply with certification standards. In this paper, the boundedness problem is addressed for a comprehensive class of uncertain dynamic systems. Several common but unnecessary approximations that are typically performed to simplify the Lyapunov analysis have been avoided in this effort. This leads to a more accurate and general formulation of the bounding set for the overall closed loop system. The conditions under which boundedness can be guaranteed are carefully analyzed; additionally, the interactions between the control design parameters, the ‘Strictly Positive Realness’ condition, and the shape and dimensions of the bounding set are discussed. Finally, an example is presented in which the bounding set is calculated for the neuro-adaptive control of an F/A-18 aircraft, along with a numerical study to evaluate the effect of several design parameters.     

Blind source separation with dynamic source number using adaptive neural algorithm

A difficult blind source separation (BSS) issue dealing with an unknown and dynamic number of sources is tackled in this study. In the past, the majority of BSS algorithms familiarize themselves with situations where the numbers of sources are given, because the settings for the dimensions of the algorithm are dependent on this information. However, such an assumption could not be held in many advanced applications. Thus, this paper proposes the adaptive neural algorithm (ANA) which designs and associates several auto-adjust mechanisms to challenge these advanced BSS problems. The first implementation is the on-line estimator of source numbers improved from the cross-validation technique. The second is the adaptive structure neural network that combines feed-forward architecture and the self-organized criterion. The last is the learning rate adjustment in order to enhance efficiency of learning. The validity and performance of the proposed algorithm are demonstrated by computer simulations, and are compared to algorithms with state of the art. From the simulation results, these have been confirmed that the proposed ANA performed better separation than others in static BSS cases and is feasible for dynamic BSS cases.     

An adaptive fuzzy-genetic algorithm approach for building detection using high-resolution satellite images

We propose a new approach for building detection using high-resolution satellite imagery based on an adaptive fuzzy-genetic algorithm. This novel approach improves object detection accuracy by reducing the premature convergence problem encountered when using genetic algorithms. We integrate the fundamental image processing operators with genetic algorithm concepts such as population, chromosome, gene, crossover and mutation. To initiate the approach, training samples are selected that represent the specified two feature classes, in this case “building” and “non-building”. The image processing operations are carried out on a chromosome-by-chromosome basis to reveal the attribute planes. These planes are then reduced to one hyperplane that is optimal for discriminating between the specified feature classes. For each chromosome, the fitness values are calculated through the analysis of detection and mis-detection rates. This analysis is followed by genetic algorithm operations such as selection, crossover and mutation. At the end of each generation cycle, the adaptive-fuzzy module determines the new (adjusted) probabilities of crossover and mutation. This evolutionary process repeats until a specified number of generations has been reached. To enhance the detected building patches, morphological image processing operations are applied. The approach was tested on ten different test scenes of the Batikent district of the city of Ankara, Turkey using 1 m resolution pan-sharpened IKONOS imagery. The kappa statistics computed for the proposed adaptive fuzzy-genetic algorithm approach were between 0.55 and 0.88. The extraction performance of the algorithm was better for urban and suburban buildings than for buildings in rural test scenes.     

Segmentation of DNA microarray images using an adaptive graph-based method

Microarray is a powerful tool for simultaneous study of the behaviour of thousands of genes through analysis of produced images. The correct segmentation of each `spot` of the microarray image is a critical step in the analysis of the results of an experiment. A graph-based method is proposed which automatically performs the segmentation. The performance of the algorithm is tested both on real and simulated images. The proposed algorithm successfully detected spots of different sizes and shapes under the presence of variable noise levels. The simulation results proved that the suggested approach has high segmentation accuracy.     

Automatic circle detection on digital images with an adaptive bacterial foraging algorithm

This article presents an algorithm for the automatic detection of circular shapes from complicated and noisy images without using the conventional Hough transform methods. The proposed algorithm is based on a recently developed swarm intelligence technique, known as the bacterial foraging optimization (BFO). A new objective function has been derived to measure the resemblance of a candidate circle with an actual circle on the edge map of a given image based on the difference of their center locations and radii lengths. Guided by the values of this objective function (smaller means better), a set of encoded candidate circles are evolved using the BFO algorithm so that they can fit to the actual circles on the edge map of the image. The proposed method is able to detect single or multiple circles from a digital image through one shot of optimization. Simulation results over several synthetic as well as natural images with varying range of complexity validate the efficacy of the proposed technique in terms of its final accuracy, speed, and robustness.     

Stochastic sampling design using a multi-objective genetic algorithm and adaptive neural networks

This paper presents a novel multi-objective genetic algorithm (MOGA) based on the NSGA-II algorithm, which uses metamodels to determine optimal sampling locations for installing pressure loggers in a water distribution system (WDS) when parameter uncertainty is considered. The new algorithm combines the multi-objective genetic algorithm with adaptive neural networks (MOGA–ANN) to locate pressure loggers. The purpose of pressure logger installation is to collect data for hydraulic model calibration. Sampling design is formulated as a two-objective optimization problem in this study. The objectives are to maximize the calibrated model accuracy and to minimize the number of sampling devices as a surrogate of sampling design cost. Calibrated model accuracy is defined as the average of normalized traces of model prediction covariance matrices, each of which is constructed from a randomly generated sampling set of calibration parameter values. This method of calculating model accuracy is called the ‘full’ fitness model. Within the genetic algorithm search process, the full fitness model is progressively replaced with the periodically (re)trained adaptive neural network metamodel where (re)training is done using the data collected by calling the full model. The methodology was first tested on a hypothetical (benchmark) problem to configure the setting requirement. Then the model was applied to a real case study. The results show that significant computational savings can be achieved by using the MOGA–ANN when compared to the approach where MOGA is linked to the full fitness model. When applied to the real case study, optimal solutions identified by MOGA–ANN are obtained 25 times faster than those identified by the full model without significant decrease in the accuracy of the final solution.     

Operation algorithm of adaptive network-based fuzzy control system for a jib crane

In this paper, we propose a data processing algorithm for the working parameters of a jib crane. The system simulates the actions of a human operator during load positioning. The algorithm of control actions generation plays the main role in operation of the automatic control system.     

Estimating neural networks-based algorithm for adaptive cachereplacement

In this paper, we propose an adaptive cache replacement scheme based on the estimating type of neural networks (NN's). The statistical prediction property of such NN's is used in our work to develop a neural network based replacement policy which can effectively identify and eliminate inactive cache lines. This would provide larger free space for a cache to retain actively referenced lines. The proposed strategy may, therefore, yield better cache performance as compared to the conventional schemes. Simulation results for a wide spectrum of cache configurations indicate that the estimating neural network based replacement scheme provides significant performance advantage over existing policies.     

Feedforward neural network and adaptive network-based fuzzy inference system in study of power lines

Over the past several decades, concerns have been raised over the possibility that the exposure to extremely low frequency electromagnetic fields from power lines may have harmful effects on human and living organisms. This paper presents novel approach based on the use of both feedforward neural network (FNN) and adaptive network-based fuzzy inference system (ANFIS) to estimate electric and magnetic fields around an overhead power transmission lines. An FNN and ANFIS used to simulate this problem were trained using the results derived from the previous research. It is shown that proposed approach ensures satisfactory accuracy and can be a very efficient tool and useful alternative for such investigations.     

图像脉冲噪声的自适应长距离相关迭代滤波法

提出一种基于局部极值噪声检测的自适应长距离相关迭代滤波算法.该算法首先采用局部极值法进行噪声检测,然后在一定的搜寻范围内计算信号点与噪声点的背景均方误差值,并以该背景均方误差值为基础采用自适应加权法进行滤波,最后将这一滤波过程进行迭代计算.实验结果表明,该算法滤波效果优于传统的滤波算法,它可以有效地去除图像中的脉冲噪声,并较好地保持图像细节信息,在噪声密度很大的情况下也表现出很好的滤波性能.     

A neural network-based approach for an assembly cell control

In several robotics applications, the robot must interact with the workspace, and thus its motion is constrained by the task. In this case, pure position control will be ineffective since forces appearing during the contacts must also be controlled. However, simultaneous position and force control called hybrid control is then required. Moreover, the nonlinear plant dynamics, the complexity of the dynamic parameters determination and computation constraints makes more difficult the synthesis of control laws. In order to satisfy all these constraints, an effective hybrid force/position approach based on artificial neural networks for multi-inputs/multi-outputs systems is proposed. This approach realizes, simultaneously, an identification and control of systems, and it is implemented according to two phases: At first, a neural observer is trained off-line on the basis of the data acquired during contact motion, in order to realize a smooth transition from free to contact motion. Then, an online learning of the neural controller is implemented using neural observer parameters so that the closed-loop system maintains a good performance and compensates for uncertain/unknown dynamics of the robot and the environment. A typical example on which we shall focus is an assembly task. Experimental results on a C5 links parallel robot demonstrate that the robot's skill improves effectively and the force control performances are satisfactory, even if the dynamics of the robot and the environment change.     

一种基于自适应遗传算法的神经网络学习算法

结合遗传算法与梯度下降法优点,提出了一种训练神经网络权值的混合优化算法,同时能够优化网络的结构.首先利用全局搜索能力可靠的遗传算法,采用递阶编码方案和自适应变异概率,同时优化网络的权值和结构,在进化结束时,能够寻到全局最优点附近的点.在遗传算法搜索结果的基础上,利用局部寻优能力较强的梯度下降法,从此点出发,进行局部搜索,最终达到网络的训练目标.与单一的遗传算法或者梯度下降法比较而言,混合优化算法的收敛速度明显提高.