In this paper, an alternative type reduction approach based on information combination with crisp interval operations (crisp interval operation type reduction [CIOTR]) is presented to avoid the time-consuming difficulty in the type reduction procedure for the mainly used interval-valued membership functions (MFs). Multiinformation is considered to be included in the primary MF of an interval-valued fuzzy set (FS). Interval operations are utilized to combine the multiinformation. With the integrated information, an interval-valued FS is reduced to be a simple FS. The proposed type reduction approach (CIOTR) is also applied to the design of a hybrid interval-valued fuzzy sliding controller (HFSCIV) for a double-pendulum-and-cart system. The simulation results are included to show the effectiveness of the designed HFSCsIV using the CIOTR.

Recently, a lossless vector quantisation (VQ) index compression algorithm, called the coding tree assignment scheme with improved search‐order coding (CTAS‐ISOC) algorithm, has been proposed to enhance the coding efficiency of the original SOC by exploiting the correlations of the neighbouring blocks using the left‐pair and upper‐pair patterns in the index domain. This study proposes a modified coding tree assignment scheme (MCTAS) to further improve the coding performance of CTAS‐ISOC by the dynamic index table coding (DITC). The DITC technique exploits the correlation of neighbouring index pairs not in the original vector‐quantised index map, but in the temporarily constructed index table that has been classified and updated for each current index. The searching matched index in a previously qualified index table results in low‐time complexity. In addition, the identical index table can be regenerated in the reconstruction process of the index map at the decoder end. Experimental results show the time complexity of MCTAS is more efficient than that of CTAS‐ISOC. Moreover, the proposed MCTAS apparently reduces the bit rate in comparison with conventional VQ and some popular lossless index coding schemes, such as index searching algorithm with index associated list, adaptive index coding scheme, SOC and so on.

In this paper, we propose a novel possibilistic fuzzy matching strategy with invariant properties, which can provide a robust and effective matching scheme for two sets of iris feature points. In addition, the nonlinear normalization model is adopted to provide more accurate position before matching. Moreover, an effective iris segmentation method is proposed to refine the detected inner and outer boundaries to smooth curves. For feature extraction, the Gabor filters are adopted to detect the local feature points from the segmented iris image in the Cartesian coordinate system and to generate a rotation-invariant descriptor for each detected point. After that, the proposed matching algorithm is used to compute a similarity score for two sets of feature points from a pair of iris images. The experimental results show that the performance of our system is better than those of the systems based on the local features and is comparable to those of the typical systems.

In this paper, the interval type-2 fuzzy controllers (FC(IT2)s) are approximated using the fuzzy ratio switching type-1 FCs to avoid the complex type-reduction process required for the interval type-2 FCs. The fuzzy ratio switching type-1 FCs (FC(FRST1)s) are designed to be a fuzzy combination of the possible-leftmost and possible-rightmost type-1 FCs. The fuzzy ratio switching type-1 fuzzy control technique is applied with the sliding control technique to realize the hybrid fuzzy ratio switching type-1 fuzzy sliding controllers (HFSC(FRST1)s) for the double-pendulum-and-cart system. The simulation results and comparisons with other approaches are provided to demonstrate the effectiveness of the proposed HFSC(FRST1)s.

Path planning is an important research topic in the autonomous robot applications. In this paper, a path is obtained by optimizing a cost function which is composed of the costs of the path length and the desired behavior. It is assumed that the desired behavior of the robot is the wall following with obstacle avoidance. First, a map of an operational environment is constructed. The cost of the desired behavior can then be assigned for the map. After that, a path can be obtained based on the behavior cost and the path length to the target. In order to operate on a fine grid map with an acceptable computation time, a multi-resolution scheme is used to acquire the optimal path. Initially, a path is obtained on a coarse resolution. Then the final path on the original fine resolution can be found within a band containing the initial path. Simulation results show that the proposed approach can find the optimal path efficiently.

In this paper, a novel fuzzy-sliding and fuzzy-integral-sliding controller (FSFISC) is designed to position the yaw and pitch angles of a twin-rotor multi-input-multi-output system (TRMS). With the coupling effects, which are considered as the uncertainties, the highly coupled nonlinear TRMS is pseudodecomposed into a horizontal subsystem and a vertical subsystem (VS). The proposed FSFISC consists of a fuzzy-sliding controller and an FISC for the horizontal and the VSs, respectively. The reaching conditions and the stability of the TRMS with the proposed controller are guaranteed. Simulation results are included to indicate that TRMS with the presented FSFISC can greatly alleviate the chattering effect and remain robust to the external disturbances. In addition, the performance comparisons with the proportional-integral-differential (PID) approach using a modified real-value-type genetic algorithm are provided to show that the FSFISC has better performance in the aspects of error and control indexes.

In this paper, an adaptive fuzzy switched swing-up and sliding controller (AFSSSC) is proposed for the swing-up and position controls of a double-pendulum-and-cart system. The proposed AFSSSC consists of a fuzzy switching controller (FSC), an adaptive fuzzy swing-up controller (FSUC), and an adaptive hybrid fuzzy sliding controller (HFSC). To simplify the design of the adaptive HFSC, the double-pendulum-and-cart system is reformulated as a double-pendulum and a cart subsystem with matched time-varying uncertainties. In addition, an adaptive mechanism is provided to learn the parameters of the output fuzzy sets for the adaptive HFSC. The FSC is designed to smoothly switch between the adaptive FSUC and the adaptive HFSC. Moreover, the sliding mode and the stability of the fuzzy sliding control systems are guaranteed. Simulation results are included to illustrate the effectiveness of the proposed AFSSSC.

In this paper, a novel clustering method in the kernel space is proposed. It effectively integrates several existing algorithms to become an iterative clustering scheme, which can handle clusters with arbitrary shapes. In our proposed approach, a reasonable initial core for each of the cluster is estimated. This allows us to adopt a cluster growing technique, and the growing cores offer partial hints on the cluster association. Consequently, the methods used for classification, such as support vector machines (SVMs), can be useful in our approach. To obtain initial clusters effectively, the notion of the incomplete Cholesky decomposition is adopted so that the fuzzy c-means (FCM) can be used to partition the data in a kernel defined-like space. Then a one-class and a multiclass soft margin SVMs are adopted to detect the data within the main distributions (the cores) of the clusters and to repartition the data into new clusters iteratively. The structure of the data set is explored by pruning the data in the low-density region of the clusters. Then data are gradually added back to the main distributions to assure exact cluster boundaries. Unlike the ordinary SVM algorithm, whose performance relies heavily on the kernel parameters given by the user, the parameters are estimated from the data set naturally in our approach. The experimental evaluations on two synthetic data sets and four University of California Irvine real data benchmarks indicate that the proposed algorithms outperform several popular clustering algorithms, such as FCM, support vector clustering (SVC), hierarchical clustering (HC), self-organizing maps (SOM), and non-Euclidean norm fuzzy c-means (NEFCM). © 2009 Wiley Periodicals, Inc.4

The k-nearest neighbor classifier is a simple and appealing approach to classification problems. It expects the class conditional probabilities to be locally constant, and suffers from bias in high dimensional situation. Using a locally adaptive metric becomes crucial in order to keep class conditional probabilities close to uniform, thereby minimizing the bias of estimates. A technique that computes a locally flexible metric by means of the decision boundaries of support vector machines (SVMs) is proposed. Then the modified neighborhoods can be shrunk in directions orthogonal to these decision boundaries and elongated parallel to the boundaries. Thereafter, any neighborhood-based classifier can use the modified neighborhoods.

In this paper, an efficient feature extraction algorithm based on optimized Gabor filters and a relative variation analysis approach is presented for iris recognition. The Gabor filters are optimized by tuning the parameters with the particle swarm optimization method. Moreover, a sequential filter scheme is developed to determine the number of filters in the optimal Gabor filter bank. In the preprocessing step, the lower part of the iris image is unwrapped and normalized to a rectangular block which is then decomposed by the optimal Gabor filters. After that, a simple encoding method is adopted to generate a compact iris code. Experimental results show that the performance of the proposed method is encouraging and comparable to those of the existing iris recognition systems.

In this paper, a systematic two-stage design method for adaptive fuzzy controllers (TSAFC's) is presented. The proposed control scheme has low computational complexity. Moreover, the exact mathematical model of the plant to be controlled is not required. The fuzzy controller under consideration is based on the proportional-derivative fuzzy control scheme and triangular membership functions. In the design procedure, the domain intervals of the input and output variables are selected with a heuristic approach to minimize the cost function under the constraint of tolerable overshoots in the response curve. A learning scheme is then proposed to automatically adjust the parameters in the fuzzy controller to reduce the error of the system. Simulations are included to demonstrate the effectiveness of the proposed method.

Face recognition is a popular research topic in the biometric identification area. Face detection and localization is the most important pre-processing module of a face recognition system. The purpose of face detection is to search and localize the positions of faces in an image in the varied background. In this paper, we propose a face detection system which combines the eigenface algorithm with the neural network. The candidate face region is examined to see if it is a face image block by analyzing the geometrical distribution of the edges in the region. In our experiments, the proposed face detection system has a high face detection rate and less false detections.

In this paper, a method is proposed for the segmentation of color images using a multiresolution-based signature subspace classifier (MSSC) with application to psoriasis images. The essential techniques consist of feature extraction and image segmentation (classification) methods. In this approach, the fuzzy texture spectrum and the two-dimensional fuzzy color histogram in the hue-saturation space are first adopted as the feature vector to locate homogeneous regions in the image. Then these regions are used to compute the signature matrices for the orthogonal subspace classifier to obtain a more accurate segmentation. To reduce the computational requirement, the MSSC has been developed. In the experiments, the method is quantitatively evaluated by using a similarity function and compared with the well-known LS-SVM method. The results show that the proposed algorithm can effectively segment psoriasis images. The proposed approach can also be applied to general color texture segmentation applications.

A robust complexity reduced proportional-integral-derivative (PID)-like fuzzy controllers is designed for a plant with fuzzy linear model. The plant model is described with the expert's linguistic information involved. The linguistic information for the plant model is represented as fuzzy sets. In order to design a robust fuzzy controller for a plant model with fuzzy sets, an approach is developed to implement the best crisp approximation of fuzzy sets into intervals. Then, Kharitonov's Theorem is applied to construct a robust fuzzy controller for the fuzzy uncertain plant with interval model. With the linear combination of input variables as a new input variable, the complexity of the fuzzy mechanism of PID-like fuzzy controller is significantly reduced. The parameters in the robust fuzzy controller are determined to satisfy the stability conditions. The robustness of the designed fuzzy controller is discussed. Also, with the provided definition of relative robustness, the robustness of the complexity reduced fuzzy controller is compared to the classical PID controller for a second-order plant with fuzzy linear model. The simulation results are included to show the effectiveness of the designed PID-like robust fuzzy controller with the complexity reduced fuzzy mechanism.

A new design approach of an adaptive fuzzy terminal sliding mode controller for linear systems with mismatched time-varying uncertainties is presented in this paper. A fuzzy terminal sliding mode controller is designed to retain the advantages of the terminal sliding mode controller and to reduce the chattering occurred with the terminal sliding mode controller. The sufficient condition is provided for the uncertain system to be invariant on the sliding surface. The parameters of the output fuzzy sets in the fuzzy mechanism are adapted on-line to improve the performance of the fuzzy sliding mode control system. The bounds of the uncertainties are not required to be known in advance for the presented adaptive fuzzy sliding mode controller. The stability of the fuzzy control system is also guaranteed. Moreover, the chattering around the sliding surface in the sliding mode control can be reduced by the proposed design approach. Simulation results are included to illustrate the effectiveness of the proposed adaptive fuzzy terminal sliding mode controller.

In this paper, a flexible complexity reduced design approach for PID-like fuzzy controllers is proposed. With the linear combination of input variables as a new input variable, the complexity of the fuzzy mechanism of PID-like fuzzy controllers is significantly reduced. However, the performance of the complexity reduced fuzzy PID controller may be degraded since the degree of freedom is decreased by the combination of input variables. To alleviate the drawback and improve the performance of the complexity reduced PID-like fuzzy controller, a flexible complexity reduced design approach is introduced in which the functional scaling factors are heuristically generated. Since the functional scaling factors are heuristically created, they can be easily adjusted for the flexible complexity reduced PID-like fuzzy controller without a priori knowledge of the exact mathematical model of the plant. Moreover, heuristic scaling factors are implemented as functionals. Therefore, the complexity of the flexible PID-like fuzzy controller will not be increased. Further, the stability of the fuzzy control system with a flexible complexity reduced PID-like fuzzy controller is discussed. Finally, the simulation results are also included to show the effectiveness of the PID-like fuzzy controller designed with the flexible complexity reduced approach.

Many common foundations exist between neural networks and fuzzy inference systems in terms of their mathematical models and system structures. This paper explores such a rich synergy and uses it to form the basis for a unifying framework under which fuzzy logic processing and neural networks may be integrated to achieve more robust information processing. It in turn leads to a family of hierarchical fuzzy neural networks (FNNs) which incorporate an adaptive and modular design of neural networks into the basic fuzzy logic systems. Several important models which are critical to the development of the the hierarchical FNN family are studied. We demonstrate how existing unsupervised and supervised learning strategies can be an integral part of a fuzzy processing framework. In addition, hierarchical structures involving both expert modules and class modules are incorporated into the FNNs. Also presented are some promising application examples.

In this paper, an approach of designing adaptive fuzzy controllers is presented to systematically develop efficient and effective rules for fuzzy controllers. The proposed fuzzy controllers are first designed with two basic fuzzy if-then rules. Then according to the design requirements of the fuzzy control system, new fuzzy if-then rules are inserted into the rule-base structure of the fuzzy controller. Initially the inserted fuzzy rules are redundant in the sense that the resultant input-output mapping of the fuzzy rules remains intact. After that the parameters of the membership functions for the fuzzy sets of the newly added fuzzy rules are trained on-line to minimize predefined cost functions. Thus, efficient fuzzy controllers can be systematically designed. Simulations for linear, nonlinear, and delayed systems are provided to show the effectiveness of the proposed approach.

In this paper, a methodology to reduce the complexity of a robust controller based on fuzzy if-then rules is proposed. The motivation and the design of this complexity-reduced fuzzy controller are presented. This fuzzy controller with the triangular membership functions and fuzzy partition methods used here leads to a region-wise linear fuzzy controller (RLFC). The properties of the region-wise linear fuzzy controllers are discussed and the reasons why they in general perform better than the PD controllers are also provided. And the simulation results based on a second order plant are included to show that the region-wise linear fuzzy controller outperforms the PD controller. We also show that the region-wise linear fuzzy controller and original fuzzy controller have similar performances.

We describe a coding scheme based on principal component analysis to compress medical images. The region of interest (tissue region) is first located. The background area can then be coded as simple models. In this situation the compression ratio can be quite high. As for the region of interest, more sophisticated algorithm will be required to achieve a high compression ratio and preserve necessary information for diagnosis at the same time. We assume that the medical images from the same modality will exhibit similar statistics. This suggests that principal component analysis will be a good candidate for the block transform coding. And it will be unnecessary to store the principal eigenvectors for each images since this information can be calculated and stored in advance. Adaptive scheme will then be used to select proper basis for transform coding. Using this scheme, the peak signal to noise ratio can reach 49.81 dB with a compression ratio of 58.5.

This paper proposes a face recognition system based on decision-based neural networks (DBNN). The DBNN adopts a hierarchical network structure with nonlinear basis functions and a competitive credit-assignment scheme. The face recognition system consists of three modules. First, a face detector finds the location of a human face in an image. Then an eye localizer determines the positions of both eyes to help generate size-normalized, reoriented, and reduced-resolution feature vectors. (The facial region proposed contains eyebrows, eyes, and nose, but excluding mouth. Eye-glasses will be permissible.) The last module is a face recognizer. The DBNN can be effectively applied to all the three modules. The DBNN based face recognizer has yielded very high recognition accuracies based on experiments on the ARPA-FERET and SCR-IM databases. In terms of processing speeds and recognition accuracies, the performance of DBNN is superior to that of multilayer perceptron (MLP). The training phase for 100 persons would take around one hour, while the recognition phase (including eye localization, feature extraction, and classification using DBNN) consumes only a fraction of a second (on Sparc10).

Supervised learning networks based on a decision-based formulation are explored. More specifically, a decision-based neural network (DBNN) is proposed, which combines the perceptron-like learning rule and hierarchical nonlinear network structure. The decision-based mutual training can be applied to both static and temporal pattern recognition problems. For static pattern recognition, two hierarchical structures are proposed: hidden-node and subcluster structures. The relationships between DBNN's and other models (linear perceptron, piecewise-linear perceptron, LVQ, and PNN) are discussed. As to temporal DBNN's, model-based discriminant functions may be chosen to compensate possible temporal variations, such as waveform warping and alignments. Typical examples include DTW distance, prediction error, or likelihood functions. For classification applications, DBNN's are very effective in computation time and performance. This is confirmed by simulations conducted for several applications, including texture classification, OCR, and ECG analysis.

An SVD approach to the so-called structure-from-motion problem was proposed by Tomasi and Kanade [1992]. The present paper extends the original motion-shape-estimation (MSE) to the multi-camera-multi-target case. The multi-target MSE problem is: given a sequence of 2D video images of multiple moving targets, the problem is to track the 3D motion of the targets and reconstruct their 3D shapes. This is further extended to multi-camera-multi-target MSE, with potential application to the 3D occlusion problem. After collection of feature points (FPs), which are sequentially tracked by a video system, the SVD may be applied to a measurement matrix formed by the FPs. The distribution of singular values would first reveal the information about the number of objects at hand. Then, using an algebraic-based subspace clustering method, the FPs may be mapped onto their corresponding objects. Thereafter, the motion and shape may be estimated from a matrix factorization.>

Several competition-based methods for classification are compared. Special attention is paid to subspace methods which are based on computing the projections of the patterns on the principal component vectors of the correlation matrices that span the pattern subspaces. A decision learning rule which updates the correlation matrices can be used to adjust the class boundary and improve the performance of the classification. A learning subspace method is proposed, and some other classification methods are reviewed. In this comparison, all of the methods are applied to a texture classification problem and the performance results are presented.>