Information Sciences - Applications最新文献

Thresholding plays an important role in image processing. To select a suitable threshold requires some criteria on which to base the selection. A criterion of maximum fuzzy entropy is developed for selecting the threshold. In this algorithm, the degree of ambiguity in an image is measured by the entropy of a fuzzy set. The threshold is selected by maximizing the fuzzy entropy of the image. The effectiveness of the algorithm is demonstrated for different bandwidths of the membership functions using noisy and vague microscopic-slide breast cancer images. The results show that this method is useful for breast cancer detection. Moreover, this method can be applied to a wide range of image processing applications.

This paper deals with a fuzzy drive expert system for an auto-cruise car. This system consists of four rule sets such as environment recognition rules, driving control rules, learning-evaluation rules, and management meta-rules. The structure of those rules is hierarchical. The fuzzy drive expert system is structured with live units; the distance extraction and image processing unit, the environment recognition unit, the control unit, the learning-evaluation unit, the I/O unit, the knowledge rule base, and the man-machine interface.

Each unit drives the fuzzy production rules which are described by sentence and symbols based on the if-then type format. Antecedent parts and consequent parts of those rules include the fuzzy words such as big, positive, wide, short, and so on. On the basis of the recognized result of environment, the control unit manipulates the steering and the throttle valve (or fuel injectors, brake pressure) for direction control and speed control of vehicle.

The vehicle drive controls on the straight road and the corner are simulated on the digital computer. The overtaking control, the tracking control, and the avoidance control of obstacles are successful and smoothable.

This paper describes a fuzzy logic anti-slip system for Heavy Duty Off Road Vehicles. The anti-slippage system is based on distributed detection and local/global fuzzy control of slippage within an interconnected system of mechatronic wheel motors.

Within this paper, the system layout, implementation, sensor data preprocessing, slip detection, four different local/global and test results under real working conditions are described. The system developed may serve as an example for successfully applied fuzzy control for a competitive (soon) commercially available product.

In the field of spatial information systems, a primary need is to develop a sound theory of topological relationships between spatial objects. A category of formal methods for representing topological relationships is based on point-set theory. In this paper, a high level calculus-based method is compared with such point-set methods. It is shown that the calculus-based method is able to distinguish among finer topological configurations than most of the point-set methods. The advantages of the calculus-based method are the direct use in a calculus-based spatial query language and the capability of representing topological relationships among a significant set of spatial objects by means of only five relationship names and two boundary operators.

This paper discusses the design of a dedicated hardware fuzzy inference system based on the generalized modus ponens (GMP) inference rule, the “mini” fuzzy implication function, the “max-min” fuzzy composition rule, and the center of gravity (COG) defuzzification method. As a special case, the implementation of a two-input (antecedents), one-output (consequent) fuzzy inference system based on the general design idea was implemented on a single field programmable gate array (FPGA) chip with an extra EPROM for defuzzification. Circuitry details and performance data are included. Additional issues such as extendibility and adaptability are also discussed.

Four benchmark examples for evaluating neural network performance are considered. The performance of the higher order moment neural array, HOMNA, on these benchmarks is explored. Comparable results for backprop networks and ARTMAP networks are available in the literature. It is shown that HOMNA trains faster and gives equivalent or better performance than either of these two alternative neural formats.

This paper presents a technique to classify images that have been elongated or contracted. The problem is formulated using conventional regular moments. It is shown that the conventional regular moment-invariants remain no longer invariant when the image is scaled unequally in the x- and y-directions. A method is proposed to form moment-invariants that do not change under such unequal scaling and shifting. By combining moments based on the theory of algebraic invariants, some of the features become rotation invariant. Results of computer simulations for images are also included, verifying the validity of the method proposed. The performance of a neural network to classify scaled, shifted, and rotated binary images is also reported.

A concept of a fuzzy control system that adjusts itself upon the error propagation between the real system and the model is introduced. Different possible error outcomes are defined, namely, right (left) error, right (left) missed error, rule lost error, symmetry lost error, right (left) shift error, and nonresponsive error. An application to the truck backer-upper control problem is presented to corroborate the theory and demonstrate the utility of the approach to fuzzy control.

This paper describes the design of a neural network based labeled object identification system, to be used for product classification at the final inspection stage of an IBM personal computer manufacturing line. The objective was to design and identification system using existing equipment that would provide robust and accurate classification, as well as a simple means for adding new product models to the system. In the first stage of the identification system, an image of the product is obtained, and the region containing the label is segmented from the rest of the image. Preprocessing operations are performed to extract the region of interest from the segmented image. Normalized and preprocessed images of the labels are compressed using a fully-connected back-propagation autoencoder network. Features extracted in this manner are used as inputs to a Learning Vector Quantization (LVQ) network, trained to classify the labels. The system so designed is shown to satisfy the primary requirements of a typical industrial classification system.

As the cognitive processes of natural language understanding and generation are better understood, it is becoming easier, nowadays, to perform machine translation. In this paper we present our work on machine translation from Arabic to English and French, and illustrate it with a fully operational system, which runs on PC compatibles with Arabic/Latin interface. This system is an extension of an earlier system, whose task was the analysis of the natural language Arabic. Thanks to the regularity of its phrase structures and word patterns, Arabic lends itself quite naturally to a Fillmore-like analysis. The meaning of a phrase is stored in a star-like data structure, where the verb occupies the center of the star and the various noun sentences occupy specific peripheral nodes of the star. The data structure is then translated into an internal representation in the target language, which is then mapped into the target text.