Showing papers in "Neural Computing and Applications in 2005"
TL;DR: A thorough review of the commonly used data fusion frameworks is presented together with important factors that need to be considered during the development of an effective data fusion problem-solving strategy.
Abstract: This paper reviews the potential benefits that can be obtained by the implementation of data fusion in a multi-sensor environment. A thorough review of the commonly used data fusion frameworks is presented together with important factors that need to be considered during the development of an effective data fusion problem-solving strategy. A system-based approach is defined for the application of data fusion systems within engineering. Structured guidelines for users are proposed.
214 citations
TL;DR: Investigation of the efficacy of cross-entropy and square-error objective functions used in training feed-forward neural networks to estimate posterior probabilities shows cross-ENTropy has significant, practical advantages over squared-error.
Abstract: This paper investigates the efficacy of cross-entropy and square-error objective functions used in training feed-forward neural networks to estimate posterior probabilities. Previous research has found no appreciable difference between neural network classifiers trained using cross-entropy or squared-error. The approach employed here, though, shows cross-entropy has significant, practical advantages over squared-error.
194 citations
TL;DR: Simulations have shown that the Qubit model solves learning problems with significantly improved efficiency as compared to the classical model, and it is suggested that the improved performance is due to the use of superposition of neural states and theUse of probability interpretation in the observation of the output states of the model.
Abstract: Neural networks have attracted much interest in the last two decades for their potential to realistically describe brain functions, but so far they have failed to provide models that can be simulated in a reasonable time on computers; rather they have been limited to toy models. Quantum computing is a possible candidate for improving the computational efficiency of neural networks. In this framework of quantum computing, the Qubit neuron model, proposed by Matsui and Nishimura, has shown a high efficiency in solving problems such as data compression. Simulations have shown that the Qubit model solves learning problems with significantly improved efficiency as compared to the classical model. In this paper, we confirm our previous results in further detail and investigate what contributes to the efficiency of our model through 4-bit and 6-bit parity check problems, which are known as basic benchmark tests. Our simulations suggest that the improved performance is due to the use of superposition of neural states and the use of probability interpretation in the observation of the output states of the model.
118 citations
TL;DR: A back propagation neural network model has been developed for the prediction of surface roughness in turning operation and the performance of the trained neural network has been tested with experimental data, and found to be in good agreement.
Abstract: In this work, a back propagation neural network model has been developed for the prediction of surface roughness in turning operation. A large number of experiments were performed on mild steel work-pieces using high speed steel as the cutting tool. Process parametric conditions including speed, feed, depth of cut, and the measured parameters such as feed and the cutting forces are used as inputs to the neural network model. Roughness of the machined surface corresponding to these conditions is the output of the neural network. The convergence of the mean square error both in training and testing came out very well. The performance of the trained neural network has been tested with experimental data, and found to be in good agreement.
96 citations
TL;DR: An application of Taguchi’s Design of Experiments is presented, to identify the optimum setting of NN parameters in a multilayer perceptron (MLP) network trained with the back propagation algorithm.
Abstract: Neural networks have been widely used in manufacturing industry, but they suffer from a lack of structured method to determine the settings of NN design and training parameters, which are usually set by trial and error. This article presents an application of Taguchi’s Design of Experiments, to identify the optimum setting of NN parameters in a multilayer perceptron (MLP) network trained with the back propagation algorithm. A case study of a complex forming process is used to demonstrate implementation of the approach in manufacturing, and the issues arising from the case are discussed.
83 citations
TL;DR: Experimental results show that not only the fast L SSVM is faster than the standard LSSVM algorithm, but also it gives better classification performance than thestandard backpropagation multilayer perceptron network.
Abstract: In this paper, we present a new system for the classification of electrocardiogram (ECG) beats by using a fast least square support vector machine (LSSVM). Five feature extraction methods are comparatively examined in the 15-dimensional feature space. The dimension of the each feature set is reduced by using dynamic programming based on divergence analysis. After the preprocessing of ECG data, six types of ECG beats obtained from the MIT-BIH database are classified with an accuracy of 95.2% by the proposed fast LSSVM algorithm together with discrete cosine transform. Experimental results show that not only the fast LSSVM is faster than the standard LSSVM algorithm, but also it gives better classification performance than the standard backpropagation multilayer perceptron network.
81 citations
TL;DR: It is proved that the same data set, when presented to neural networks in different forms, can provide slightly varying results and also proves that when different opinions of more classifiers on the same input data are integrated within a multi-classifier system, it can obtain results that are better than the individual performances of the neural networks.
Abstract: The paper presents a neural network based multi-classifier system for the identification of Escherichia coli promoter sequences in strings of DNA. As each gene in DNA is preceded by a promoter sequence, the successful location of an E. coli promoter leads to the identification of the corresponding E. coli gene in the DNA sequence. A set of 324 known E. coli promoters and a set of 429 known non-promoter sequences were encoded using four different encoding methods. The encoded sequences were then used to train four different neural networks. The classification results of the four individual neural networks were then combined through an aggregation function, which used a variation of the logarithmic opinion pool method. The weights of this function were determined by a genetic algorithm. The multi-classifier system was then tested on 159 known promoter sequences and 171 non-promoter sequences not contained in the training set. The results obtained through this study proved that the same data set, when presented to neural networks in different forms, can provide slightly varying results. It also proves that when different opinions of more classifiers on the same input data are integrated within a multi-classifier system, we can obtain results that are better than the individual performances of the neural networks. The performances of our multi-classifier system outperform the results of other prediction systems for E. coli promoters developed so far.
68 citations
TL;DR: Simulations show that the OPA–orthogonal least squares (OPA–OLS) algorithm, which combines OPA with the OLS algorithm, results in better performance for forecasting trends, and is applied to stock price prediction.
Abstract: A novel neural-network-based method of time series forecasting is presented in this paper. The method combines the optimal partition algorithm (OPA) with the radial basis function (RBF) neural network. OPA for ordered samples is used to perform the clustering for the samples. The centers and widths of the RBF neural network are determined based on the clustering. The difference of the objective functions of the clustering is used to adjust the structure of the neural network dynamically. Thus, the number of the hidden nodes is selected adaptively. The method is applied to stock price prediction. The results of numerical simulations demonstrate the effectiveness of the method. Comparisons with the hard c-means (HCM) algorithm show that the proposed OPA method possesses obvious advantages in the precision of forecasting, generalization, and forecasting trends. Simulations also show that the OPA–orthogonal least squares (OPA–OLS) algorithm, which combines OPA with the OLS algorithm, results in better performance for forecasting trends.
50 citations
TL;DR: Classifying alert versus drowsy states in an arbitrary subject using 5-s long sequences of full-spectrum electroencephalogram (EEG) recordings showed that the MLPNN trained with the Levenberg–Marquardt algorithm was effective for discriminating the vigilance state of the subject.
Abstract: In this study, 5-s long sequences of full-spectrum electroencephalogram (EEG) recordings were used for classifying alert versus drowsy states in an arbitrary subject. EEG signals were obtained from 30 healthy subjects and the results were classified using a wavelet-based neural network. The wavelet-based neural network model, employing the multilayer perceptron (MLP), was used for the classification of EEG signals. A multilayer perceptron neural network (MLPNN) trained with the Levenberg–Marquardt algorithm was used to discriminate the alertness level of the subject. In order to determine the MLPNN inputs, spectral analysis of EEG signals was performed using the discrete wavelet transform (DWT) technique. The MLPNN was trained, cross-validated, and tested with training, cross-validation, and testing sets, respectively. The correct classification rate was 93.3% alert, 96.6% drowsy, and 90% sleep. The classification results showed that the MLPNN trained with the Levenberg–Marquardt algorithm was effective for discriminating the vigilance state of the subject.
48 citations
TL;DR: A comparative evaluation of the performance of NNs and HMMs for a TCM application is presented, which will assist the condition-monitoring community to choose a modeling method for other applications.
Abstract: Condition monitoring of machine tool inserts is important for increasing the reliability and quality of machining operations. Various methods have been proposed for effective tool condition monitoring (TCM), and currently it is generally accepted that the indirect sensor-based approach is the best practical solution to reliable TCM. Furthermore, in recent years, neural networks (NNs) have been shown to model successfully, the complex relationships between input feature sets of sensor signals and tool wear data. NNs have several properties that make them ideal for effectively handling noisy and even incomplete data sets. There are several NN paradigms which can be combined to model static and dynamic systems. Another powerful method of modeling noisy dynamic systems is by using hidden Markov models (HMMs), which are commonly employed in modern speech-recognition systems. The use of HMMs for TCM was recently proposed in the literature. Though the results of these studies were quite promising, no comparative results of competing methods such as NNs are currently available. This paper is aimed at presenting a comparative evaluation of the performance of NNs and HMMs for a TCM application. The methods are employed on exactly the same data sets obtained from an industrial turning operation. The advantages and disadvantages of both methods are described, which will assist the condition-monitoring community to choose a modeling method for other applications.
38 citations
TL;DR: Application results demonstrate that the proposed techniques can significantly improve model generalisation and perform better than aggregating all the individual networks.
Abstract: A Bayesian selective combination method is proposed for combining multiple neural networks in nonlinear dynamic process modelling. Instead of using fixed combination weights, the probability of a particular network being the true model is used as the combination weight for combining that network. The prior probability is calculated using the sum of squared errors of individual networks on a sliding window covering the most recent sampling times. A nearest neighbour method is used for estimating the network error for a given input data point, which is then used in calculating the combination weights for individual networks. Forward selection and backward elimination are used to select the individual networks to be combined. In forward selection, individual networks are gradually added into the aggregated network until the aggregated network error on the original training and testing data sets cannot be further reduced. In backward elimination, all the individual networks are initially aggregated and some of the individual networks are then gradually eliminated until the aggregated network error on the original training and testing data sets cannot be further reduced. Application results demonstrate that the proposed techniques can significantly improve model generalisation and perform better than aggregating all the individual networks.
TL;DR: The proposed neural network interpolation method is based on wavelet reconstruction, which obtains much better performance than other traditional methods and is applicable to various other related problems.
Abstract: Changing the resolution of digital images and video is needed image processing systems. In this paper, we present nonlinear interpolation schemes for still image resolution enhancement. The proposed neural network interpolation method is based on wavelet reconstruction. With the wavelet decomposition, the image signals can be divided into several time–frequency portions. In this work, the wavelet decomposition signal is used to train the neural networks. The pixels in the low-resolution image are used as the input signal of the neural network to estimate all the wavelet sub-images of the corresponding high-resolution image. The image of increased resolution is finally produced by the synthesis procedure of wavelet transform. In the simulation, the proposed method obtains much better performance than other traditional methods. Moreover, the easy implementation and high flexibility of the proposed algorithm also make it applicable to various other related problems.
TL;DR: This paper proposes a new codebook generation algorithm for image data compression using a combined scheme of principal component analysis (PCA) and genetic algorithm (GA).
Abstract: This paper proposes a new codebook generation algorithm for image data compression using a combined scheme of principal component analysis (PCA) and genetic algorithm (GA). The combined scheme makes full use of the near global optimal searching ability of GA and the computation complexity reduction of PCA to compute the codebook. The experimental results show that our algorithm outperforms the popular LBG algorithm in terms of computational efficiency and image compression performance.
TL;DR: This study presents the development and implementation of two supervised vector quantization neural networks for estimating the flank-land wear size of a twist drill and found that the FLVQ is more efficient in assessing the flank wear size than the LVQ.
Abstract: Drill wear detection and prognosis is one of the most important considerations in reducing the cost of rework and scrap and to optimize tool utilization in hole making industry. This study presents the development and implementation of two supervised vector quantization neural networks for estimating the flank-land wear size of a twist drill. The two algorithms are; the learning vector quantization (LVQ) and the fuzzy learning vector quantization (FLVQ). The input features to the neural networks were extracted from the vibration signals using power spectral analysis and continuous wavelet transform techniques. Training and testing were performed under a variety of speeds and feeds in the dry drilling of steel plates. It was found that the FLVQ is more efficient in assessing the flank wear size than the LVQ. The experimental procedure for acquiring vibration data and extracting features in the time-frequency domain using the wavelet transform is detailed. Experimental results demonstrated that the proposed neural network algorithms were effective in estimating the size of the drill flank wear.
TL;DR: It turns out, the Multi-layer perceptron does best when used without confirmation filters and leverage, while the Softmax cross entropy model and the Gaussian mixture model outperforms the Multi/Dollar perceptron when using more sophisticated trading strategies and leverage.
Abstract: Dunis and Williams (Derivatives: use, trading and regulation 8(3):211–239, 2002; Applied quantitative methods for trading and investment. Wiley, Chichester, 2003) have shown the superiority of a Multi-layer perceptron network (MLP), outperforming its benchmark models such as a moving average convergence divergence technical model (MACD), an autoregressive moving average model (ARMA) and a logistic regression model (LOGIT) on a Euro/Dollar (EUR/USD) time series. The motivation for this paper is to investigate the use of different neural network architectures. This is done by benchmarking three different neural network designs representing a level estimator, a classification model and a probability distribution predictor. More specifically, we present the Mulit-layer perceptron network, the Softmax cross entropy model and the Gaussian mixture model and benchmark their respective performance on the Euro/Dollar (EUR/USD) time series as reported by Dunis and Williams. As it turns out, the Multi-layer perceptron does best when used without confirmation filters and leverage, while the Softmax cross entropy model and the Gaussian mixture model outperforms the Multi-layer perceptron when using more sophisticated trading strategies and leverage. This might be due to the ability of both models using probability distributions to identify successfully trades with a high Sharpe ratio.
TL;DR: This paper presents an application of a radial basis support vector machine (RB-SVM) to the recognition of the sleep spindles (SSs) in electroencephalographic (EEG) signal and shows that the best performance is obtained with an RB-S VM providing an average sensitivity of 97.7%.
Abstract: This paper presents an application of a radial basis support vector machine (RB-SVM) to the recognition of the sleep spindles (SSs) in electroencephalographic (EEG) signal. The proposed system comprises of two stages. In the first stage, for feature extraction, a set of raw amplitude values, a set of discrete cosine transform (DCT) coefficients, a set of discrete wavelet transform (DWT) approximation coefficients and a set of adaptive autoregressive (AAR) parameters are calculated and extracted from signals separately as four different sets of feature vectors. Thus, four different feature vectors for the same data are comparatively examined. In the second stage, these features are then selected by a modified adaptive feature selection method based on sensitivity analysis, which mainly supports input dimension reduction via selecting the most significant feature elements. Then, the feature vectors are classified by a support vector machine (SVM) classifier, which is relatively new and powerful technique for solving supervised binary classification problems due to it’s generalization ability. Visual evaluation, by two electroencephalographers (EEGers), of 19 channel EEG records of six subjects showed that the best performance is obtained with an RB-SVM providing an average sensitivity of 97.7%, an average specificity of 97.4% and an average accuracy of 97.5%.
TL;DR: Application of radial basis neural networks (RBNN) for prediction of cavitation vortex dynamics in a Francis turbine draft tube is presented and a good agreement between power spectra and correlation functions of measured and predicted void fractions was shown.
Abstract: Application of radial basis neural networks (RBNN) for prediction of cavitation vortex dynamics in a Francis turbine draft tube is presented. The dynamics of the cavitation vortex was established by fluctuations of a void fraction in a selected region of the draft tube. The void fraction was determined by image acquisition and analysis. Pressure in the draft tube and images of the cavitation vortex were acquired simultaneously for the experiment. RBNN were used for prediction. The void fraction in the selected region of the cavitation vortex was predicted on the basis of experimentally provided pressure data. The learning set consisted of pressure – void fraction pairs. The prediction consisted in providing only the pressure. Regression coefficients r between the predicted and measured void fractions were in an interval of 0.82–0.98. A good agreement between power spectra and correlation functions of measured and predicted void fractions was shown.
TL;DR: The denoising layers, when integrated into feedforward and recurrent neural networks, were validated on three time series prediction problems: the logistic map, a rubber hardness time series, and annual average sunspot numbers.
Abstract: To avoid the need to pre-process noisy data, two special denoising layers based on wavelet multiresolution analysis have been integrated into layered neural networks. A gradient-based learning algorithm has been developed that uses the same cost function to set both the neural network weights and the free parameters of the denoising layers. The denoising layers, when integrated into feedforward and recurrent neural networks, were validated on three time series prediction problems: the logistic map, a rubber hardness time series, and annual average sunspot numbers. Use of the denoising layers improved the prediction accuracy in both cases.
TL;DR: This paper has predicted the type of pregnancy, as well as the factors influencing it using two different models and comparing them, and developed a multinomial logistic regression and a neural network based on the data and compared their results using three statistical indices: sensitivity, specificity and kappa coefficient.
Abstract: Some medical and epidemiological surveys have been designed to predict a nominal response variable with several levels. With regard to the type of pregnancy there are four possible states: wanted, unwanted by wife, unwanted by husband and unwanted by couple. In this paper, we have predicted the type of pregnancy, as well as the factors influencing it using two different models and comparing them. Regarding the type of pregnancy with several levels, we developed a multinomial logistic regression and a neural network based on the data and compared their results using three statistical indices: sensitivity, specificity and kappa coefficient. Based on these three indices, neural network proved to be a better fit for prediction on data in comparison to multinomial logistic regression. When the relations among variables are complex, one can use neural networks instead of multinomial logistic regression to predict the nominal response variables with several levels in order to gain more accurate predictions.
TL;DR: This work proposes the application of pruning in the design of neural networks for hydrological prediction, and shows that pruned models may provide better generalization than fully connected ones, thus improving the quality of the forecast.
Abstract: We propose the application of pruning in the design of neural networks for hydrological prediction. The basic idea of pruning algorithms, which have not been used in water resources problems yet, is to start from a network which is larger than necessary, and then remove the parameters that are less influential one at a time, designing a much more parameter-parsimonious model. We compare pruned and complete predictors on two quite different Italian catchments. Remarkably, pruned models may provide better generalization than fully connected ones, thus improving the quality of the forecast. Besides the performance issues, pruning is useful to provide evidence of inputs relevance, removing measuring stations identified as redundant (30–40% in our case studies) from the input set. This is a desirable property in the system exercise since data may not be available in extreme situations such as floods; the smaller the set of measuring stations the model depends on, the lower the probability of system downtimes due to missing data. Furthermore, the Authority in charge of the forecast system may decide for real-time operations just to link the gauges of the pruned predictor, thus saving costs considerably, a critical issue in developing countries.
TL;DR: Simulated networks of spiking leaky integrators are used to categorise and for Information Retrieval (IR) and the results show that congresspeople are correctly categorised 89% of the time.
Abstract: Simulated networks of spiking leaky integrators are used to categorise and for Information Retrieval (IR). Neurons in the network are sparsely connected, learn using Hebbian learning rules, and are simulated in discrete time steps. Our earlier work has used these models to simulate human concept formation and usage, but we were interested in the model’s applicability to real world problems, so we have done experiments on categorisation and IR. The results of the system show that congresspeople are correctly categorised 89% of the time. The IR systems have 40% average precision on the Time collection, and 28% on the Cranfield 1,400. All scores are comparable to the state of the art results on these tasks.
TL;DR: A feedforward multi-layer perceptron neural network structure is developed to model the nonlinear dynamic relationship between input and output of a hydro power plant connected as single machine infinite bus system.
Abstract: A feedforward multi-layer perceptron neural network structure is developed to model the nonlinear dynamic relationship between input and output of a hydro power plant connected as single machine infinite bus system. Two independent second-order neural network nonlinear auto-regressive with exogenous signal models are used in the study. The structure selection of each independent model is based on various validation tests. The optimal brain surgeon pruning strategy adopted for optimizing the neural network structure. The network performance is studied for fixed and change in operating point.
TL;DR: The proposed scheme is useful in solving the security problems that occurred in systems using the password table and verification table and allows each user to select a username and password of his/her choice.
Abstract: Information security has been a critical issue in the field of information systems. One of the key factors in the security of a computer system is how to identify the authorization of users. Password-based user authentication is widely used to authenticate a legitimate user in the current system. In conventional password-based user authentication schemes, a system has to maintain a password table or verification table which stores the information of users’ IDs and passwords. Although the one-way hash functions and encryption algorithms are applied to prevent the passwords from being disclosed, the password table or verification table is still vulnerable. In order to solve this problem, in this paper, we apply the technique of back-propagation network instead of the functions of the password table and verification table. Our proposed scheme is useful in solving the security problems that occurred in systems using the password table and verification table. Furthermore, our scheme also allows each user to select a username and password of his/her choice.
TL;DR: A hybrid neural network that is capable of incremental learning and classification of patterns with incomplete data is proposed, employed as the constituting network for pattern classification while fuzzy c-means clustering is used as the underlying algorithm for processing training as well as test samples with missing features.
Abstract: In this paper, a hybrid neural network that is capable of incremental learning and classification of patterns with incomplete data is proposed. Fuzzy ARTMAP (FAM) is employed as the constituting network for pattern classification while fuzzy c-means (FCM) clustering is used as the underlying algorithm for processing training as well as test samples with missing features. To handle an incomplete training set, FAM is first trained using complete samples only. Missing features of the training samples are estimated and replaced using two FCM-based strategies. Then, network training is conducted using all the complete and estimated samples. To handle an incomplete test set, a non-substitution FCM-based strategy is employed so that a predicted output can be produced rapidly. The performance of the proposed hybrid network is evaluated using a benchmark problem, and its practical applicability is demonstrated using a medical diagnosis task. The results are compared, analysed and quantified statistically with the bootstrap method. Implications of the proposed network for pattern classification tasks with incomplete data are discussed.
TL;DR: The radial basis function-based Gaussian mixture model (GMM) is applied to model and predict drug dissolution profiles in a time-series approach and implications of the GMM for pharmaceutical product formulation tasks are discussed.
Abstract: In this paper, the radial basis function-based Gaussian mixture model (GMM) is applied to model and predict drug dissolution profiles in a time-series approach. The Parzen-window method is embedded into the GMM for determining whether the network predictions are derived from interpolation or extrapolation of the training data. A benchmark study on time-series prediction is first used to evaluate and compare the GMM performance with those from other models. The GMM is then used to predict dissolution profiles of a matrix-controlled release theophylline pellet preparation. Performance of the GMM is assessed using the difference and similarity factors, as recommended by the United States Food and Drug Administration for dissolution profile comparison. In addition, bootstrapping is employed to estimate the confidence intervals of the network predictions. The experimental results are analyzed and compared, and implications of the GMM for pharmaceutical product formulation tasks are discussed.
TL;DR: Two types of immune multi-agent neural networks, which have agents of macrophages, B-cells and T-cells, are described and their classification capabilities are compared.
Abstract: Many different learning algorithms for neural networks have been developed, with advantages offered in terms of network structure, initial values of some parameters, learning speed, and learning accuracy. If we train the networks only on good examples, without noise and shortage, the neural network can be trained to classify, with reasonable accuracy, target patterns or random patterns, but not both. To solve this problem, we propose a learning method of immune multi-agent neural networks (IMANNs), which have agents of macrophages, B-cells and T-cells. Each agent employs a different type of neural network. Because the agents work cooperatively and competitively, IMANNs can automatically classify the training dataset into some subclasses. In this paper, two types of IMANNs are described and their classification capabilities are compared. In order to verify the effectiveness of our proposed method, we used two datasets: the dataset of the MONK’s problem (as a traditional classification problem) and a dataset from a medical diagnosis problem (hepatobiliary disorders).
TL;DR: A general approach to embed deep and shallow knowledge in neural network models for fault diagnosis by abduction, using neural sites for logical aggregation of manifestations and faults, is proposed.
Abstract: In real systems, fault diagnosis is performed by a human diagnostician, and it encounters complex knowledge associations, both for normal and faulty behaviour of the target system. The human diagnostician relies on deep knowledge about the structure and the behaviour of the system, along with shallow knowledge on fault-to-manifestation patterns acquired from practice. This paper proposes a general approach to embed deep and shallow knowledge in neural network models for fault diagnosis by abduction, using neural sites for logical aggregation of manifestations and faults. All types of abduction problems were considered. The abduction proceeds by plausibility and relevance criteria multiply applied. The neural network implements plausibility by feed-forward links between manifestations and faults, and relevance by competition links between faults. Abduction by plausibility and relevance is also used for decision on the next best test along the diagnostic refinement. A case study on an installation in a rolling mill plant is presented.
TL;DR: This paper proposes to use hierarchical neural networks with local recurrent connectivity to solve the task of localizing a human face in an image not only in unambiguous situations, but also in the presence of complex backgrounds, difficult lighting, and noise.
Abstract: One of the major tasks in some human–computer interface applications, such as face recognition and video telephony, is to localize a human face in an image. In this paper, we propose to use hierarchical neural networks with local recurrent connectivity to solve this task not only in unambiguous situations, but also in the presence of complex backgrounds, difficult lighting, and noise. The networks are trained using a database of gray-scale still images and manually determined eye coordinates. They are able to produce reliable and accurate eye coordinates for unknown images by iteratively refining initial solutions. Because the networks process entire images, there is no need for any time-consuming scanning across positions and scales. Furthermore, the fast network updates allow for real-time face tracking. In this case, the networks are trained using still images that move in random directions. The trained networks are able to accurately track the eye positions in the test image sequences.
TL;DR: This special issue comprises six papers selected from more than 350 presented at the Seventh International Conference on Knowledge-Based Intelligent Information and Engineering Systems (KES2003), 3–5 September 2003, University of Oxford, UK, which showcase a range of applications of neural networks to real-world problems, e.g. bioinformatics, medical diagnosis, fault diagnosis and image and pattern recognition.
Abstract: Neural networks have been applied with tremendous success to a wide range of real-world problems. The authors are continuing to be delighted to present the advantages offered by various neural networks in developing applications of great practical value where other techniques employed have failed. Neural networks have proven their practical applicability to almost any type of research or application area where learning for further recognition, prediction or classification tasks is concerned. This special issue comprises six papers selected from more than 350 presented at the Seventh International Conference on Knowledge-Based Intelligent Information and Engineering Systems (KES2003), 3–5 September 2003, University of Oxford, UK. The annual KES Conference has become a traditional and prestigious world forum for the presentation of developments relating to a wide spectrum of intelligent techniques and applications, with neural networks being among the most popular topics. The selected papers showcase a range of applications of neural networks to real-world problems, e.g. bioinformatics, medical diagnosis, fault diagnosis and image and pattern recognition. In addition to an interesting range of applications, the papers included in this issue come up with new techniques and approaches on hot topics in the area of neural networks, such as quantum neural networks, neural-network-based immune and multi-agent systems and multi-classifier systems. Today human–computer interfaces require advanced image processing systems. One difficult task in image processing is the localisation of the human face in an image. In the first paper of this issue, S. Behnke from the University of Freiburg, Germany proposed a hierarchical neural network architecture with local recurrent connectivity for face localisation and tracking. The proposed network is able to recognise human faces in the presence of noise, complex backgrounds or difficult lighting. Moreover, the network can be used for realtime face tracking, e.g. for accurately localising human faces in moving images. A delicate issue in machine learning is how to deal with incomplete datasets. C.P. Lim, M.M. Kuan and R.F. Harrison present a FAM-FCM hybrid neural network model for pattern classification when incomplete data sets are available. The fuzzy ARTMAP (FAM) network, which integrates fuzzy techniques with adaptive resonance theory (ART) neural networks, is the main part of the system and is employed to classify the input samples under various conditions. The fuzzy cmeans (FCM) based clustering strategies involved in this research are responsible for handling samples with missing features. The authors demonstrate the practical applicability of the FAM-FCM hybrid system in a medical diagnosis problem where, usually, the datasets available for making the diagnosis are incomplete. Over the years, researchers in the field of neural networks have proposed various types of neural networks, and have shown their limitations in a variety of real-world applications. Recently, quantum computing has been seen as a good solution for improving the abilities of neural networks. The third paper, by N. Kouka, N. Matsui, H. Nishimura and F. Pepper, introduces a quantum-computing-based neuron model, called Qubit. The quantum neural networks built using Qubit are compared with classical feed-forward neural networks, as well as with complex neural networks, in some pattern recognition problems. The performance of their quantum neural networks is reported to be better. The fourth paper is concerned with an application of neural networks to the fascinating field of bioinformatics. R. Ranawana and V. Palade from the University of Oxford, UK present a neural network multi-classifier system for gene recognition in DNA sequences of the Escherichia Coli (E. Coli) microorganism. Using various encoding techniques for input data, more neural network classifiers are trained using the same dataset, and V. Palade (&) Oxford University, UK
TL;DR: A new PCA learning algorithm based on cascade recursive least square (CRLS) neural network is proposed that can guarantee the network weight vector converges to an eigenvector associated with the largest eigenvalue of the input covariance matrix globally.
Abstract: Principal component analysis (PCA) by neural networks is one of the most frequently used feature extracting methods. To process huge data sets, many learning algorithms based on neural networks for PCA have been proposed. However, traditional algorithms are not globally convergent. In this paper, a new PCA learning algorithm based on cascade recursive least square (CRLS) neural network is proposed. This algorithm can guarantee the network weight vector converges to an eigenvector associated with the largest eigenvalue of the input covariance matrix globally. A rigorous mathematical proof is given. Simulation results show the effectiveness of the algorithm.