An automatic defect classification system (ADC) for use in visual inspection of semiconductor wafers is introduced. The methods of extracting the defect features based on the human experts' knowledge, with their correlations with the defect classes are elucidated. As for the classifier, Hyperellipsoid Clustering Network (HCN) which is a layered network model employing second order discrimination borders in the feature space, is introduced. In the experiments using a collection of defect images, the HCNs are compared with the conventional multilayer perceptron networks. There, it is shown that the HCN's adaptive hyperellipsoidal discrimination borders are more suited for the problem. Also, the cluster encapsulation by the hyperellipsoidal border enables to determine rejection classes, which is also desirable when the system will be in actual use. The HCN with rejection achieves, an overall classification rate of 75% with an error rate of 18%, which can be considered equivalent to those of the human experts.

A novel neural network architecture for image texture classification is introduced. The proposed model (Kernel Modifying Neural Network: KM Net) which incorporates the convolution filter kernel and the classifier in one, enables an automated texture feature extraction in multichannel texture classification through the modification of the kernel and the connection weights by the backpropagation-based training rule. The first layer units working as the convolution kernels are constrained to be an array of Gabor filters, which achieves a most efficient texture feature localization. The following layers work as a classifier of the extracted texture feature vectors. The capability of the KM Net and its training rule is verified using a basic problem on a synthetic texture image. In addition, the possibilities of applying the KM Net to natural texture classification and biological tissue classification using an ultrasonic echo image have been tried.

Control of the pulse collision point on an axon, prior to the propagation time measurement facilitates the detection of the non-uniformity of an axon, similar to a TDR on transmission lines. The measurement requires the placement of electrodes only on the extremities of the axon.

This paper presents an application of Complex-Valued Associative Memory Model(CAMM) for image processing. An image association system applying CAMM, combined with a 2-dimensional discrete Fourier transform (2-D DFT) process is proposed. Discussed are how a gray level image can be expressed using CAMM, and the image association that can be performed by CAMM. In the proposed system, input images are transformed to phase matrices and the image association can be performed by making use of the phase information. Practical examples are also presented.

In order to improve the efficiency of the feature extraction of backpropagation (BP) learning in layered neural networks, model switching for changing the function model without altering the map is proposed. Model switching involves map preserving reduction of units by channel fusion, or addition of units by channel installation. For reducing the model size by channel fusion, two criteria for detection of the redundant channels are addressed, and the local link weight compensations for map preservation are formulated. The upper limits of the discrepancies between the maps of the switched models are derived for use as the unified criterion in selecting the switching model candidate. In the experiments, model switching is used during the BP training of a layered network model for image texture classification, to aid its inefficiency of feature extraction. The results showed that fusion and re-installation of redundant channels, weight compensations on channel fusion for map preservation, and the use of the unified criterion for model selection are all effective for improved generalization ability and quick learning. Further, the possibility of using model switching for concurrent optimization of the model and the map will be discussed.

In laparoscopic surgery, the lack of tactile sensation and 3D visual feedback make it difficult to identify the position of a blood vessel intraoperatively. An unintentional partial tear or complete rupture of a blood vessel may result in a serious complication; moreover, if the surgeon cannot manage this situation, open surgery will be necessary. Differentiation of arteries from veins and other structures and the ability to independently detect them has a variety of applications in surgical procedures involving the head, neck, lung, heart, abdomen, and extremities. We have used the artery's pulsatile movement to detect and differentiate arteries from veins. The algorithm for change detection in this study uses edge detection for unsupervised image registration. Changed regions are identified by subtracting the systolic and diastolic images. As a post-processing step, region properties, including color average, area, major and minor axis lengths, perimeter, and solidity, are used as inputs of the LVQ (Learning Vector Quantization) network. The output results in two object classes: arteries and non-artery regions. After post-processing, arteries can be detected in the laparoscopic field. The registration method used here is evaluated in comparison with other linear and nonlinear elastic methods. The performance of this method is evaluated for the detection of arteries in several laparoscopic surgeries on an animal model and on eleven human patients. The performance evaluation criteria are based on false negative and false positive rates. This algorithm is able to detect artery regions, even in cases where the arteries are obscured by other tissues.

Segmenting the images obtained from magnetic resonance imaging (MRI) is an important process for visualization of the human soft tissues. For the application of MR, we often have to introduce a reasonable segmentation technique. Neural networks may provide us with superior solutions for the pattern classification of medical images than the conventional methods. For image segmentation with the aid of neural networks of a reasonable size, it is important to select the most effective combination of secondary indices to be used for the classification. In this paper, we introduce a vector quantized class entropy (VQCCE) criterion to evaluate which indices are effective for pattern classification, without testing on the actual classifiers. We have exploited a newly developed neural tree classifier for accomplishing the segmentation task. This network effectively partitions the feature space into subregions and each final subregion is assigned a class label according to the data routed to it. As the tree grows on, the number of training data for each node decreases, which results in less weight update epochs and decreases the time consumption. The partitioning of the feature space at each node is done by a simple neural network; the appropriateness of which is measured by newly proposed estimation criterion, i. e. the measure for assessment of neuron (MAN). It facilitates the obtaining of a neuron with maximum correlation between a unit's value and the residual error at a given output. The application of this criterion guarantees adopting the best-fit neuron to split the feature space. The proposed neural classifier has achieved 95% correct classification rate on average for the white/gray matter segmentation problem. The performance of the proposed method is compared to that of a multilayered perceptron (MLP), the latter being widely exploited network in the field of image processing and pattern recognition. The experiments show the superiority of the introduced method in terms of less iterations and weight up dates necessary to train the neural network, i. e. lower computational complexity; as well as higher correct classification rate.

We introduce a concept of regularization into Genetic Algorithms (GAs). Conventional GAs include no explicit regularizing operations. However, the regularization is very effective in solving ill-posed problems. So, we propose a method of regularization to apply GAs to ill-posed problems. This regularization is a kind of consensus operation among neighboring individuals in GAs, and plays the role of `smoothing the solution. ' Our method is based on the evaluation of macroscopic fitness, which is a new fitness criterion. Conventional fitness of an individual in GAs is defined only from the phenotype of the individual, whereas the macroscopic fitness of an individual is evaluated from the phenotypes of the individual and its neighbors. We tested our regularizing operation by means of experiments with an elastic image mapping problem, and showed the effectiveness of the regularization.

This paper presents a novel feature extraction algorithm based on particle swarms for processing hyperspectral imagery data. Particle swarm optimization, originally developed for global optimization over continuous spaces, is extended to deal with the problem of feature extraction. A formulation utilizing two swarms of particles was developed to optimize simultaneously a desired performance criterion and the number of selected features. Candidate feature sets were evaluated on a regression problem. Artificial neural networks were trained to construct linear and nonlinear models of chemical concentration of glucose in soybean crops. Experimental results utilizing real-world hyperspectral datasets demonstrate the viability of the method. The particle swarms-based approach presented superior performance in comparison with conventional feature extraction methods, on both linear and nonlinear models.

This paper investigates the modeling of non-linearity on the generation of the single trial evoked potential signal (s-EP) by means of using a mixed radial basis function neural network (M-RBFN). The more emphasis is put on the contribution of spontaneous EEG term to s-EP signal. The method is based on a nonlinear M-RBFN neural network model that is trained simultaneously with the different segments of EEG/EP data. Then, the output of the trained model (estimator) is a both fitted and reduced (optimized) nonlinear model and then provide a global representation of the passage dynamics between spontaneous brain activity and poststimulus periods. The performance of the proposed neural network method is evaluated using a realistic simulation and applied to a real EEG/EP measurement.

An electronic model of the coronary vessel consisting of resistor, capacitor and Field Effect Transistor (FET) is proposed in order to perform a dynamic simulation of the left coronary circulation and to clarify its mechanisms. Based on this model, an equivalent circuit of the coronary circulation is constructed that is divided into subepicardial and subendocardial layers and consists of segments of artery, arteriole, capillary, venule and vein for both the layers. In this simulation, the observed flow waveform of the left main artery showed dominance of flow in diastole as compared to that in systole. In epicardium, inverse venous flow was observed in early systole. These simulated waveforms are similar to those in real left coronary circulation observed by physiological and clinical researchers. Among all the segments of intramyocardium, only the venules were found to possess a time-varying resistance characteristics. From the results of this study, it is considered that the combination of resistance and capacitance of the vessel acts as an integrator and a differentiator for blood pressure and intramyocardial pressure, respectively and that the effects of integration of blood pressure and differentiation of intramyocardial pressure play a very important role in determining the factors influencing the left coronary circulation.

We propose an algorithm for image decomposition based on Hadamard functions, realized by answer-in-weights neural network, which has simple architecture and is explored with steepest decent method. This scheme saves memory consumption and it converges fast. Simulations with least mean square (LMS) and absolute mean (AM) errors on a 128128 image converge within 30 training epochs.

Neural network pruning is a technique to obtain a fully functional subset of a redundant network for the efficiency of computation. A new method to prune a redundant three-layered neural network by means of neural element fusion" is introduced. In contrast to conventional pruning techniques that remove unimportant portions of the network, our method fuses a pair of hidden layer units so that features accumulated in both units are preserved as possible. The pair of hidden layer units to be fused is chosen by evaluating a firing similarity. This similarity measure also informs when the pruning should be stopped. The fusing method was compared with well known unit removing" methods on computer simulations. The results show that our fusing method considerably reduces the error increase due to the pruning, even in subminimal networks where conventional methods are ineffective. This enables to cut down the total cost of computation to reach the minmal network configuration.

In this paper, our aim is to extract real-time movement-related potentials, especially readiness-potentials, from EEGs with a small number of scalp electrodes. We proposed a method composed of independent component analysis (ICA), dipole tracing (DT) and scalp Laplacian methods. The proposed method shows a good real-time RP extraction capability from a single-trial of movement by means of the selection of EEGs with high reliability based on the DT and the improvement of the spatial resolution of the scalp potentials based on the scalp Laplacian.

A new neural network for locating a source by integrating data from a number of sensors is considered. The network gives a solution for inverse problems using a back-propagation algorithm with the architecture to get the solution in the inter-layer weights in a coded form Three different physical quantities are applied to the network, since the scheme has three independent ports; an input port, a tutorial port and an answer port. Our architecture is useful to estimate z" in the problem whose structure is y=f(x,z) where y is the observed data, x is the sensor position and z is the source location. The network integrates the information obtained from a number of sensors and estimates the location of the source. We apply the network to two problems of location estimation: the localization of the active nerves from their evoked potential waveforms and the localization of objects from their echoes using an active sonar system.