摘要:

AbstractThis paper describes a high‐speed parallel clustering method that is suitable for use as a preprocessor for motion picture compression and coding. In the parallel clustering process, plural pixels are sampled in one frame of an image. Clusters are gathered according to their characteristics, among them, colors, motion vectors, etc., as a way to organize clusters or modify cluster‐parameters in parallel. This operation is repeated until the frame is segmented optimally by a set of clusters. We propose the subframe parallel approach in order to enhance the parallelism of the clustering method and avoid load‐imbalance in parallel tasks. The effectiveness of this approach is also presented and it is shown that a real‐time clustering of motion picture images is available in mos

ISSN:0882-1666    

DOI:10.1002/scj.4690271401

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1996

数据来源: WILEY

摘要:

AbstractThis paper proposes a novel networking architecture on a distributed processing environment (DPE) for various multimedia services suitable to high‐speed networks and for the flexible and rapid introduction services in an open networking environment. Furthermore, the proposed architecture is realized as a testbed. This architecture has the following features: (1) the use of DPE for the rapid introduction of service to users and for the flexible management of multimedia network resources; (2) the definition of the service session and connection management for efficient managing functions of the session and connection; and (3) the introduction of any special service with the use of an Ad Hoc service manager. The interfaces scheme between a service session manager that is implemented on a DPE platform and an Ad Hoc service manager is an important aspect in an open networking environment. A CSCW application is given as an example of an Ad Hoc service manage

ISSN:0882-1666    

DOI:10.1002/scj.4690271402

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1996

数据来源: WILEY

摘要:

AbstractA responsive system, which serves to unify a fault‐tolerant system and real‐time system, must have flexible real‐time fault recovery capability. Of the multiprocessor systems in which the number of active processors changes dynamically due to failures and repairs of processors, those with the reconfiguration capability to isolate faulty processors from the system and reconnect repaired processors to the system, in addition to the ability to process tasks in real‐time, are examples of responsive systems. Normal task processing operations may be interrupted by reconfiguration and tasks requiring real‐time processing may miss their deadline as a result. A modeling method on a stochastic reward net is proposed in this paper to represent the failure‐repair behavior of a system and changes in external environments (including fluctuations in the task arrival rate). This method explicitly considers the possibility of missed deadlines. As a result of performance evaluation using the proposed model, it is shown that system performance can be improved by controlling the timing of the system reconfiguration, thus minimizing the effects of system reco

ISSN:0882-1666    

DOI:10.1002/scj.4690271403

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1996

数据来源: WILEY

摘要:

AbstractCombinatorial optimization problems are solved here by using a neural network that includes maximum functions in the energy functions. Because the energy function is bilinear in the original Hopfield neural network, the objective optimization problem must be formulated in bilinear form. It is shown that an energy function having maximum functions can be formulated; in addition, the neural network corresponding to the energy function is able to solve optimization problems. The example used here solves the maximum satisfiability problem and the results are compared with two simple algorithms. An oscillatory unit already in use is applied to the proposed neural network and the effect is examined. Neural networks can handle solutions for a logical function having 5—20 literals and 100—100,000 nodes. The neural network with oscillatory units presented here puts out better soluti

ISSN:0882-1666    

DOI:10.1002/scj.4690271404

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1996

数据来源: WILEY

摘要:

AbstractThis paper proposes a method of speech recognition in a nonstationary noisy environment, combining the parallel HMMs and the spectral subtraction. In the proposed method, a set of hypothesis is generated with respect to the combination of the speech and the noise that can produce the observed data by a series of subtraction processes. Using HMMs prepared separately for the speech and the noise, the probabilities of occurrence are calculated.The 100‐word recognition in the noisy environment in an ordinary car running in an urban area, is defined as the task in the experiment. Comparative experiments, are made for the proposed method, the ordinary spectral subtraction method and other parallel HMM methods. Then, the effectiveness of the proposed method is verifie

ISSN:0882-1666    

DOI:10.1002/scj.4690271405

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1996

数据来源: WILEY

摘要:

AbstractIn the hand‐eye system, the contour of the object is usually extracted through the Hough transform of the image in order to determine the shape and orientation of an object based on the input image. When this procedure is actually applied, however, the contour cannot be extracted with sufficient accuracy, due to the effect of noise in the image or a quantization error in the Hough transform. From such a viewpoint, this paper proposes a method to extract the line segment with high accuracy and no overlap by estimating the quantization error in the Hough transform. The effectiveness of this method is demonstrated using an actual exampl

ISSN:0882-1666    

DOI:10.1002/scj.4690271406

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1996

数据来源: WILEY

摘要:

AbstractThis paper considers the neural network in which the initial values for the weights and the bias are given by random numbers as in usual cases. The results of BP learning in networks composed of unipolar units having an activity range from 0 to 1 and networks with bipolar units with a range from −0.5 to 0.5 are compared.When the input space is large, the separation hyperplane at the outset of learning passes near the center of the input space in the bipolar case, while that in the unipolar case passes near the vertex. Because of this property, the number of separation hyperplanes that effectively separate the input spaces of the layers during the updating or realization of the solution is larger in the bipolar case than in the unipolar case. The difference between the two becomes more remarkable with the increase of size. As a result of simulation, it is verified that the learning by the bipolar network gives better convergence for a wider range of initial values than the learning by the unipolar network when the network is large. It is shown also that the kinds of solution obtained by the unipolar network tend to be deviate

ISSN:0882-1666    

DOI:10.1002/scj.4690271407

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1996

数据来源: WILEY

摘要:

AbstractThis paper proposes a method of rotating a three‐dimensional image on a screen by using a two‐dimensional pointer (e.g., a mouse cursor) to make medical diagnoses easier. The apparatus rotates the three‐dimensional image by directly clicking and dragging areas of interest with a pointer so that a user (e.g., a busy doctor) can examine the image from any number of angles without a break in his/her concentration. Experiments show that the proposed method is faster by 20 percent and simpler and easier to use without special training as compared with a conventional method (in which a mouse cursor points in a numerical viewing angle direc

ISSN:0882-1666    

DOI:10.1002/scj.4690271408

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1996

数据来源: WILEY

摘要:

AbstractThis paper proposes a method to reconstruct a space‐variant blurred image by using the wavelet transform, which is an orthogonal transform producing space‐variant frequency data. A blurred image is divided into multiple resolutions by using a wavelet transform. As each resolution depends on the position of the image, this is corrected by applying a coefficient determined by a PSF parameter (the half‐maximum width) so that the position‐dependent frequency data are restored. The PSF in the blurred image is estimated to realize blind deconvolution. The method has been tested by a computer simulation and by a CCD camera using real images. The results show that the PSF in the blurred images can be estimated accurately when the images contain little noise. When the resolution of each blurred image is corrected by the estimated PSF parameter, it has been confirmed that the restoration was carried out depending on the degree of b

ISSN:0882-1666    

DOI:10.1002/scj.4690271409

出版商:Wiley Subscription Services, Inc., A Wiley Company    

年代:1996

数据来源: WILEY