摘要:

AbstractIt is recognized as necessary to have a software development environment that can simultaneously store software products and maintain their detailed relationships. There exists the cut&paste model based on attribute grammars, which is effective in maintaining consistency among software products and in automatically maintaining the derived values. This model has a problem however, namely the description of the dynamical aspect of the software development environment is difficult. From such a viewpoint, an object‐oriented attribute grammar type computation model OOAG has been proposed, which introduces the concept of the message passing into attribute grammars. The OOAG is defined with an extension and reorganization in regard to the dynamical description functions. Then, the problem in the cut&past model are discussed. As an example of the problems, it is pointed out that description is difficult for the case where the point of referral by the user and the point of modification are different. Such a situation is described by OOAG, demonstrating that OOAG can remedy the problem in the cut&paste mode

ISSN:0882-1666    

DOI:10.1002/scj.4690270701

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

年代:1996

数据来源: WILEY

摘要:

AbstractDistance maps of binary images contain, for each pixel, the distance between that pixel and the pixel of value 0 closest to it. They are useful for a variety of machine‐vision applications. This paper presents a simple and efficient algorithm for computing the Euclidean distance maps. The algorithm runs inO(N2)time for anN × Nbinary image, and it also works for other distances that appear in machine‐vision applications, such as city block and chessboard. Its parallel version runs inO(N2/p)time withp(1 ≤p≤N) pr

ISSN:0882-1666    

DOI:10.1002/scj.4690270702

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

年代:1996

数据来源: WILEY

摘要:

AbstractVarious versions of the majority‐voting classification method have been proposed in recent years as a strategy for improving classification performance. This method generates multiple decision trees from training examples and performs majority voting of classification results from these decision trees in order to classify test examples. In this method, however, since the target concept is represented in multiple decision trees, its readability is poor. This property makes it ineffective in knowledge‐base construction.To enable the majority‐voting classification method to be applied to knowledge‐base construction, this paper proposes a simplification method that converts the entire majority‐voting classifier into compact disjunctive normal form (DNF) formulas. A significant feature of this method is the use of binary decision diagrams (BDDs) as internal expressions in the conversion process to achieve high‐speed simplification. A problem that must be addressed here is the BDD input variable ordering scheme. This paper proposes an ordering scheme based on the order of variables in the decision trees.The simplification method has been applied to several real‐world data sets of the Irvine Database and to data from medical diagnosis domain. It was found that the description size of the majority‐voting classifier after simplification was on the average from 1.2 to 2.7 times that of a single decision tree and was less than one‐third the size of a majority‐voting classifier before simplification. Therefore, the method is effective in reducing the description size and should be applicable to the knowledge acquisition process. Using the input variable ordering scheme proposed here, high‐speed simplification of several seconds to several tens of seconds is achieved on a Sun SPARC

ISSN:0882-1666    

DOI:10.1002/scj.4690270703

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

年代:1996

数据来源: WILEY

摘要:

AbstractWe study the applicability and evaluate the decryption fidelity of a spread spectrum image encryption method, for which decryption is accomplished with easy operations, and which makes possible the protection of secrecy even in hardcopy form. Decryption in the spread spectrum encryption method is based on the recognition of the dark‐pixel region in the picture's background noise. To evaluate the applicability of this image encryption method, we employ measureBn, called the boldfacedness, which estimates the boldness of the original image. This is essentially the area of the dark‐pixel region normalized by the length of the path surrounding it. We relate the intuitive visibility of several kinds of images to the boldfacedness, and confirm its validity. To estimate its applicability, we composed a questionnaire concerning the recognition of simple rectangular images in background noise. The results showed that for these simple rectangular images aBnvalue down to about 0.023 is adequate. Next, since we often want to transmit images of documents in facsimile, we evaluated the fidelity of document images based on the size and type of several kinds of characters. To perform this evaluation we took as input Group 3 facsimile, with the resolution of all document images as a consistent 200 dpi. It became clear that, for document images, the letter‐recognition threshold value ofBnis in the neighborhood of 0.015. This is equivalent to the case both for 16‐point Mincho‐tai characters and for 10‐point standard characters, verifying the applicability of this encryption method for general doc

ISSN:0882-1666    

DOI:10.1002/scj.4690270704

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

年代:1996

数据来源: WILEY

摘要:

AbstractAn automatic speech recognition method is proposed using the Markov model in which a large number of states is adopted to model the transitional characteristics of speech more accurately. Unlike the traditional HMM, the feature vector of this proposed model is considered to be the parameter of the state of the Markov model.First, the transition‐probability of the state and the symbol output probability are estimated on the assumption that they are represented by multidimensional normal density functions of the feature vector. The optimal time sequence is determined by using DP‐matching.Next, the multistate model is obtained by quantizing the feature vector space and calculating (or sampling) the value of the probability density function at each code vector (at multipoints). The probability distribution of the multistate Markov model is adapted by the retraining method to be more accurate in distribution. The resulting recognizer is evaluated on a vocabulary of English four‐digit numerals. The multistate model has attained a recognition score of 98.2 percent, i.e., 1.6 percent higher than that of a five‐state traditional HMM. The processing time for the reference pattern generation is one‐fiftieth that o

ISSN:0882-1666    

DOI:10.1002/scj.4690270705

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

年代:1996

数据来源: WILEY

摘要:

AbstractThis paper proposes a high‐speed algorithm to extract regions for gray‐scale image based on the evaluation of the features of each connected component in each gray level. In the proposed method, first, the input gray‐scale image is divided into region components having the same gray level, and the image is represented using an adjacent and a gray‐level depth which symbolizes the connective relations between the detected region elements. Then, by searching this graph, the relationship between connected regions of different gray‐levels, and the features of each region in each gray level are detected and represented by a tree. Using this representation, the optimal thresholds to extract regions which maximize evaluations of their features can be determined individually, with the amount of computation proportional only to the number of initial c

ISSN:0882-1666    

DOI:10.1002/scj.4690270706

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

年代:1996

数据来源: WILEY

摘要:

AbstractA method has been investigated in which the video content description is handled by a computer and is applied to the video database, as a means of desktop video production. In the course of video production, it often happens that the description of the interpretation of the video content by the video producer is retained in the form of a document. Such a description is considered most important to describe the video, but there have been only a few efforts at handling the description by electronic means. This paper proposes a method to analyze the description style of the scenario or the raw memorandum by the on‐site producer, and the corresponding video component scene description model is constructed. The application to the video database is discussed, in order to demonstrate the usefulness of the model. An experiment to extract the layered structure is presented using the video in the actual broadcast program. The effectiveness of the structurization technique and the feasibility of the application of the descriptive model to the video database are demonstrated. An experimentally constructed video browser is also discusse

ISSN:0882-1666    

DOI:10.1002/scj.4690270707

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

年代:1996

数据来源: WILEY

摘要:

AbstractUntil now, the successive state‐splitting algorithm (SSS) has been proposed as an automatic generation algorithm for the hidden Markov network (HMnet), which can be used as a high‐performance efficient context‐dependent model. SSS is a splitting‐type algorithm in which the HMnet is detailed only by splitting states.On the other hand, the “merge‐type” algorithm also is proposed, where the model is constructed while merging the parameters based on the partial similarity of the model parameters. The two methods have the common goal, but take completely different approaches. Consequently, they should complement each other.From such a viewpoint, this paper introduces the mechanism of the “merge‐type” technique into SSS as the “splitting‐type” algorithm. In other words, the “state‐splitting and merging” algorithm is proposed, where a higher‐performance automatic generation of HMnet is realized by complementing the defects of the two methods. An HMnet actually is constructed using 25 Japanese phoneme samples. By various evaluation experiments, the performance of the p

ISSN:0882-1666    

DOI:10.1002/scj.4690270708

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 for building a system which automatically images the components of a human face. The method is useful in any number of different perspectives. It is assumed that facial appearance depends on the camera angle and perspective as well as individual facial characteristics. Appearances of a face are classified based on the applicable possibilities of the vision algorithms. Two‐dimensional models of different facial appearances are constructed and the algorithms based on these models are applied in parallel to each scene in a top‐down mode. An appearance model consists of agents, each of which has simple functions. The results of the analysis are re‐combined with their assumptions based on the Dempster‐Shafer probability model. Experiments using actual facial images confirm the usefulness of the proposed

ISSN:0882-1666    

DOI:10.1002/scj.4690270709

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

年代:1996

数据来源: WILEY