From: rick@cs.arizona.edu (Rick Schlichting) Subject: INDUSTRY: Kahaner Report: Inst Image Elec Eng of Japan, Date: 12 Aug 92 20:21:49 GMT Message-ID: <20468@optima.cs.arizona.edu> Crossposted from comp.research.japan [Dr. David Kahaner is a numerical analyst on sabbatical to the Office of Naval Research-Asia (ONR Asia) in Tokyo from NIST. The following is the professional opinion of David Kahaner and in no way has the blessing of the US Government or any agency of it. All information is dated and of limited life time. This disclaimer should be noted on ANY attribution.] [Copies of previous reports written by Kahaner can be obtained using anonymous FTP from host cs.arizona.edu, directory japan/kahaner.reports.] To: Distribution From: David K. Kahaner US Office of Naval Research Asia (From outside US): 23-17, 7-chome, Roppongi, Minato-ku, Tokyo 106 Japan (From within US): Unit 45002, APO AP 96337-0007 Tel: +81 3 3401-8924, Fax: +81 3 3403-9670 Email: kahaner@cs.titech.ac.jp Re: Inst Image Electronics Eng of Japan, 20th Ann mtg., 9-11 June 92, Tokyo 7 August 1992 This file is named "iieej.92" ABSTRACT. Titles and abstracts presented at the Institute of Image Electronics Engineers of Japan, 9-11 June 1992, Tokyo Japan This small meeting brought together about 100 people to listen to almost 50 contributed papers from 10 countries. A variety of topics were discussed, but mainly focused on Image recognition and understanding Printing/fax technology Multi media coding In addition, a third day was devoted to an advanced session on Document processing and digital imaging at which about 20 additional papers were presented. Papers numbered 1-44 were delivered in Japanese, 45-47 and A1 - A19 in English. A Proceedings (296pp) containing all the papers was published by Institute of Image Electronics Engineers of Japan No. 505, 2-21-27, Shimo-Meguro, Meguro-ku, Tokyo 153 JAPAN Tel: +81 3-3490-2940; Fax: +81 3-3490-3810 Below we list the titles and authors of each paper. In addition, English abstracts, when these were available are given. (* identifies the speaker.) Please contact me or the IIEEJ for copies of these papers or other information. 1 Mathematical Morphology and the Structural Analysis of Kanji *S. Go, W. Chit, S. Shu and H. Tominaga (Waseda Univ.) 2 Coding Techniques for High Definition Text Images Using Drawing Strokes and The Continuity of Blocks *H. Iizuka, O. Nakamura and T. Minami (Kogakuin Univ.) 3 An Objective Evaluation of Picture Quality for Code Images *A. Sakurai and K, Kirisawa (Seikei Univ.) 4 Initial Value Setting of Adaptive Binary Arithmetic Coding for Text/File Compression S. Kino, *K. Maruyama and K. Kondo (Mitsubishi Electric Corp.) 5 Quality Estimation of Extremely Small Printed Characters *T. Kobayasi, Y. Numagami, O. Nakamura and T. Minami (Kogakuin Univ.) - Organized Session - Image Recognition and Understanding 6 Toward Enhanced Character Recognition Technology *T. Wakahara (Inst. for Posts and Telecommunications Policy) 7 Image Recognition Technology for Video Editing Support *H. Ueda (Hitachi, Ltd.) 8 Japanese Address Reader-Sorter *Y. Nakamura (Toshiba Corp.) - General Session 9 Intelligent Image Coding and Its Application (Invited Speech) *Y. Hatori (KDD) 10 Identification of Intersections Using Image Matching *S. Misono and M. Aoki (Seikei Univ.) 11 Interpretation of Optical Flow Using Neural Network Learning M. Seki and *M. Miyauchi (Sanno College) 12 Body Motion Tracking of Gymnastic Exercise *A. Mitsui and M. Aoki (Seikei Univ.) 13 Collation of Facial Images Based on Isodensity Maps *K. Ueno, O. Nakamura and T. Minami (Kogakuin Univ.) 14 A Study on Improvement of Reconstruction Accurucy in Stereoscopy Using Isodensity Maps *M. Fujishima, O. Nakamura and T. Minami (Kogakuin Univ.) 15 Thermal Printing System for 16dpm Half-Tone Printing *J. Iwaki, N. Inoue, K. Sasaki and N. Oshima (Toshiba Corp.) 16 Development of Personal Plane Paper Facsimile HI-TWIN *S. Matsumoto, M. Watanabe, S. Ohara and H. Nakazawa (Hitachi, Ltd.) 17 Color Matching Technique for Thermal Transfer Color Printer *T. Sato, F. Takeda, Y.Yamamoto and M. Watanabe (Hitachi, Ltd.) 18 Full Color Facsimile System Using Ion Flow Printing *M. Omodani, M. Ohta, M. Fujita and Y. Hoshimo (NTT) 19 Seamless Image Connection Technique in Image Input Using Multi-Sensor *K. Uehira (NTT) 20 A Consideration of Transmission System for Facsimile Broadcasting *H. Seno and I. Misono (NHK) 21 Processing Method of Multiply-Free DCT *J. Hara and N. Murayama (Ricoh Co.) 22 Design of Parallel Filter Bank and Optimization of Its Filter Coefficients *N. Komatsu, J. Katto and Y. Yasuda (Univ. of Tokyo) 23 A Reduction Method of Markov States with Vector Quantization *N. Inoue (Fujitsu Ltd.) and Y. Yasuda (Univ. of Tokyo) 24 A Quautization Table Design Method Based on Visual Chacteristics *D. Ando and H. Ibraki (NTT) 25 Synthesis Facial Expression Based on Motion Analysis of Isodensity Maps *Y. Hishinuma, I. So, O. Nakamura and T. Minami (Kogakuin Univ.) 26 Chaos Image Coding with Edge-Fill Conversion *H. Taniguchi, K. Suzuki and N. Murayama (Ricoh Co.) 27 A Study on IFS Estimation (without Using RIA) *H. Sumiyoshi, Y. Suzuki and A. Miyauchi (Musashi Inst. of Tech.) 28 Unreversible Encoding of Digitized Line-Drawing Images by Generalized Chain Codes *A. Watanabe and I. Masuda (Secom Co., Ltd.) - Organized Session - Multimedia Coding 29 Directions of MPEG Standardization and Image Coding Technology *Y Katayama (GC Technology Corp.) 30 Multimedia/Hypermedia Standardization *H Onoe (Victor Co. of Japan Ltd.) and Y. Yamagishi (Sony Corp.) 31 Representation Function for Multimedia/Hypermedia Standardization *T Nakagawa (NTT), W, Kameyama (Waseda Univ.) and H. Onoe (Victor Co. of Japan Ltd.) - General Session 32 An Algorithm on Generation Simplified Map According to Document Architecture for Map *K, Takahashi, K. Abe, W. Kameyama and H. Tominaga (Waseda Univ.) 33 Integration of Image and Voice Data Using JPEG Algorithm Y. Yamamoto, *K. Tanaka, Y. Nakamura and K. Matsui (The National Defense Academy) 34 Document Syntax of HI-CAPTAIN HG *M. Morinaga, K. Takikawa and R. Suzuki (NTT) 35 Fundental Experiments on Retrieval of Image Database Using Image Indexes M. Ishigame (Akita Univ.) and *M. Ishi (Tokyo Inst. of Tech.) 36 Vector Fields in the Color Image and It's Application to the Edge Detections *K. Shinohara and T. Minami (Kogakuin Univ.) - Kenote Address Mobile Communications in next generation M. Kuwabara (NTT) - General Session 37 Hivision Still Picture Disk (Invited Speech) - Focus on Standardization for Software Exchange - *K. Murakami (Meisei Univ.) 38 A Density Reduction Method for Binary Image Including Pseudo Gray Scale *S. Aoki, S. Ohuchi, K. Imao and W. Yamada (Ricoh Co., Ltd.) 39 An Adaptive Coding Method of Documents Including Character and Picture *H. Shirasawa, K. Imao and W. Yamada (Ricoh Co., Ltd.) 40 High Quality Digital Image Processing Method Using Moire Suppression and Edge Enhancement Filtering *T. Nakamura, S. Shinoda, K. Nakashima and T. Inuzuka (Hitachi, Ltd.) 41 Pixel Density Conversion Method for Halftone Images *T. Semasa, *T. Takahashi, Y. Imanaka and F. Ono (Mitsubishi Electric. Co) 42 Entropy of Bi-level Image Quantized by Minimized Average Error *M. Matsumoto, I. Ueno, T. Semasa and F. Ono (Mitsubishi Electric Corp.) 43 Evaluation of Direct Plate Printing Recorder *H. Moriwaki (NEC Engineering Ltd.), H. Endoh (NEC Co.), S. Nitta and K. Asahiria (A@i Shinbun Publishing Co.) 44 Compact, High-speed and Versatile Multi-Font Generator T. Tanaka and *T. Niwa (NHK) S. Ujihara and N. Okazaki (Matsushita Communication Ind Co.,Ltd.) [Note abstracts of remaining papers follow the listing of their titles.] 45 Arabic Document Processor: An Object Oriented Programming Approach M. N. Mikhail and K. Zaki (The American Univ. in Cairo) 46 On the Modelling of Colour Distribution in Colour Screens B. Kruse (Linkoping Univ.) 47 The NCS Colour System from Different Viewpoints B. Kruse (Linkoping Univ.) Advanced Session of IIEEJ on Document Processing and Digital Imaging A-1 A Keynote Address Document Preparation System "Yamabuki" and Its Document Style Handling Y. Ohno and Y. Yamamoto (Keio Univ., Japan) - Document recognition & Algorithms A-2 A Design of the Graphic Interface for Book Card Understanding System H. Hase, M. Yoneda, M. Sakai and T. Sasaki (Toyama Univ., Japan) A-3 Segmentation of Tables M. S. Krishnamoorthy and T. Hu (Rensselaer Polytechnic Institute, USA) A-4 Automatic Testing of Graphical Text A. Roberts, D. C. Ranyard and K. W. Brodlie (Univ. of Leeds, England) A-5 On the Recognition of Arabic Scripts Using Features Selected in the Hough Transform Space M. Fakir and C. Sodeyama (Nagaoka Univ. of Technology, Japan) A-6 A New Expression Method of Document Images Based on Maximum Matching of Bipartite Graph K. Douniwa, S. Chinveeraphan and M. Sato (Tokyo Inst of Tech, Japan) A-7 A Prototype Color Image Server J. Wiseman, L. Mather, H. Nguyen, C. Rosenberg and H. Trontelj (Hewlett-Packard Labs., USA) - Fonts A-8 Two-Dimensional Operators for Chinese Character Description M. Aoki and A. Dodo (Seikei Univ., Japan) A-9 A formalism For Hierarchical Outline-Fonts B. Stamm (Swiss Federal Institute of Technology, Switzerland) A-10 Advanced Typography P. Karow (URW Unternehmensberatung, Germany) A-11 The Design and Implementation of CCPE J. Fan (National Research Center, P.R. China) A-12 A Method of Binarization Based on Information of Character Structure N. Sun, M. Ito, J. Kudoh, K. Jayanthi, Y. Nemoto and H. Aso (Tohoku University, Japan) M. Kimura (Japan Advanced Institute of Science and Technology, East, Japan) - Database A-13 SGML-based Document Sharing System D. Oria, T. Sasaki, H. Okutomi, M. Hasunuma and K. Toyoda (Matsushita Graphics Communication System Inc., Japan) A-14 ELIS (Electronic Library System): A HyperBook System M. Y. Sung, S. J. Oh, H. G. Kang and J. S. Kim (ETRI, Korea) A-15 Technical Documentation Handling within a Computer Assisted Design Environment F. Prunet, J. Nanard, M. Nanard, T. Renoir, T. Giaconne and G. Momeret (Univ. of Montpellier, France) - System A-16 Multilingual Document Preparation Environment and its Tools Z. Ming (Univ. of Liverpool, UK) H. Ying, L. Jinxin and Z. Ning (Peking Univ., China) A-17 Arabic Word Processor: An Object Oriented Programming Approach M. N. Mikliail and K. Zaki (The American Univ. in Cairo, Egypt) A-18 Editing Facilities of Fine Multi-lingual Gray-scale Text S. Abe, M. Maekawa (Univ. of Tokyo, Japan) A-19 Audio Graphic Conference System Using Color LCD Tablet K. Yoshimi, K. Yanaka, Kazumi Komiya, T. Nomura and S. Sakai (NI7, Japan) ABSTRACTS from IIEEJ 20th Annual Meeting and Advanced Session June-9-June 11, 1992 Arabic Word Processor: An Object Oriented Programming Approach M.N. Mikhail (Computer Science Dept., The American University in Cairo, Cairo, Egypt. Fax: +202 355 7565), et al. Abstract Object Oriented programming (OOP) has proven to be the ideal environment for developing word processors. In particular, in the two most important components, namely the internal structure and the user interface. In the internal structure, a unified approach is used to represent documents as a bidirectional linked list of a document. The same structure is used to represent a document as pages and lines. The user interface is implemented using two general objects, window and button. The window could be resized, dragged, zoomed in or out, or closed. The window is linked to the document structure which would be controlled through it. Once the general button is created along with other structures such as radio buttons and check boxes, they could be called from different dialog boxes to perform different actions as the user may wish. On the Modelling of Colour Distribution in Colour Screens Bjorn Kruse (Dept. of Electrical Engineering, Linkoping University, S-581 83 Linkoping, Sweden. Fax: +46 13-28-25-99) Abstract In colour reproduction the rendition of colour is a result of the combined effect of paper and ink properties and lighting conditions. The notion of dotgain describes the process. The notion of dotgain describes the process. This paper deals with the analysis of the optical part of the dotgain. The colour distributions are viewed as 3-D surfaces as a function of varying ink coverage percentages and reflection values. It is shown how the dotgain can be analyzed by processing the tristimulus values from a scanned tint. A model is also developed for the behaviour of the distributions. A Document Preparation System "Yamabuki" and Its Document Style Handling Yoshio Ohno (Faculty of Science & Technology, Keio University, Kohoku-ku, Yokohama 223, Japan; Email: ohno@on.cs.keio.ac.jp), et al. Abstract Many people from several organizations have been developing a Japanese document preparation system called "Yamabuki" [1] for UNIX workstations. In this paper, we describe how document styles are treated in Yamabuki. Yamabuki is basically a WYSIWYG-type system, but the concept of document style is also built in this system. Therefore, the users of Yamabuki can enjoy the advantages of markup-type systems. We will discuss how we merged WYSIWYG and markup in Yamabuki. We also describe the mechanism for the novice user with little knowledge for constructing good styles. A Design of the Graphic Interface for Book Card Understanding System Hiroyuki Hase (Faculty of Engineering, Toyama University, Toyama-shi, 930, Japan. Fax: +81 764-41-8432), et al. Abstract We formerly proposed an understanding system for book cards. The aim of the system was to classify characters on a book card into several items (such as title, author, publisher and so on) for the purpose of making a book information data base effectively. The system was implemented on the main frame computer using Prolog and Fortran languages. Recently, we have switched it over from the main frame to the workstation in order to add interactive functions, and realized a graphic interface using X-window. An operator can give following commands through the interface; (1) book card input, (2) removal of noise which the system could not remove automatically, (3) confirmation and correction of recognized characters, (4) confirmation and correction of classified items, (5) registration of result, and so on. According to the correction by the operator, the system retry to get better classification results. The interface has been implemented by C programming language. In this paper, we describe the functions of the interface and the relation between the understanding system and the interface. We also give some experimental results for book cards in the library of Toyama university. Segmentation of Tables M.S. Krishnamoorthy (Dept. of Computer Science, Rensselaer Polytechnic Institute, Troy, NY 12180), et al. Abstract This paper describes an algorithm for identifying rows and columns of a table from a scanned image of a table. The paper also describes the limitation of the proposed method. The algorithms that are derived in the paper are extensions of the algorithms that are used for segmenting text document into paragraphs. Heuristic algorithms based on the number of white and black pixels are used for segmentation. Finally the paper provides experimental results for the performance of the proposed algorithms. Automatic Testing of Graphical Text A. Roberts (School of Computer Studies, University of Leeds, Leeds, LS2 9JT, England. Fax: +44 532-335-468), et al. Abstract This work is part of a project with the overall aim of reducing the human element in conformance testing of graphics software. In this paper work is described which illustrates the use of pattern classification on data retrieved by a pixel readback facility as a means of automatically testing the correctness of graphical text. The study dealt with upper and lower case letters of the Roman alphabet and even for arbitrary fonts and rotated characters a recognition success rate of around 90% was achieved. The work was done on GKS but would be readily applicable to other graphics standards. On the Recognition of Arabic Scripts Using Features Selected in the Hough Transform Space Mohamed Fakir (Dept. of Electrical Engineering, Nagaoka University of Technology, 940-21, Kamitomiokamachi, 1603-1, Nagaoka, Niigata, Japan), et al. Abstract A Method for the recognition of Arabic printed scripts entered from an image scanner is presented. After the process of smoothing, midline drift correction, and lines separation, Arabic words are segmented into characters, then characters into main body and stress marks (if any). From the character main body, features are extracted in the Hough transform space. Next characters are classified into two steps: In the first one, character main body is classified using Dynamic programming matching technique and features selected in Hough transform space. In the second one, simple topological features extracted from the geometry of the stress marks are used by a topological classifier to completely recognize the character. The method was tested on a sample set of 300 words and a recognition of 90% was obtained. A New Expression Method of Document Images Based on Maximum Matching of Bipartite Graph Ken'ichi Douniwa, Supoj Chinveeraphan (Precision and Intelligence Lab., Tokyo Institute of Technology, 4259 Nagatsuda-cho, Yokohama-shi, Kanagawa 227, Japan), et al. Abstract This paper presents a new method to express binary images, named Minimum Covering Run (MCR) expression. The proposed expression is invented by selecting suitable runs from horizontal and vertical run types which minimize the total number of runs that still represent the images exactly. It is shown that horizontal and vertical runs of binary image can be thought as partite sets of a bipartite graph. Consequently, the MCR expression that corresponds to maximum matching in the graph, can be found by using an algorithm solving this graph-theoretic problem. An efficient algorithm for the expression method is also demonstrated and the validity of stroke extraction of information in document images is shown as an application. A Prototype Color Image Server Jeanne Wiseman (Hewlett-Packard Labs, Computer Peripherals Lab, 1501 Page Mill Road, Palo Alto, CA 94304, USA. Fax: 415-857-4681), et al. Abstract We present the architecture and algorithms of a prototype Color Image Server prototype. The system consists of an Image Server program which manages compressed color images on a workstation and of a companion Image Browser client used to retrieve visually, display, print, and capture the images. The architecture combines image compression with progressive transmission to minimize retrieval times on the network while keeping the storage requirements low. A novel redundancy reduction lossy compression algorithm was developed specifically for this application where images are more often decompressed than compressed. The algorithm offers significant speed advantage for decoding over transform based techniques, traded for a small degradation in image quality and high encoding complexity. Progressive transmission is achieved by encoding the images in a 3-layer pyramid structure using this compression technique at different bit rates and in different color spaces. The last layer of the pyramid provides visually lossless encoding suitable for high quality display and printing. Two-Dimensional Operators for Chinese Character Description Masayoshi Aoki (Dept. of Electrical Engineering and Electronics, Seikei University, 3-1 Kichijoji Kitamachi-3, Musashino-shi, Tokyo 180, Japan Fax: +81 422-373871), et al. Abstract When using a word processor, the traditional Chinese character indexing system is not easy to use for a novice user or a casual user. Many characters have simple structure from a two dimensional geometrical point of view. We proposed to use not only the already existing named "Bushu", but also Chinese character itself. For the two dimensional composition operation, only three operators will be used, ie. H(horizontal), V(vertical) and A(add, insert). We applied the proposed scheme to "JIS level 1" Chinese characters (2,965 characters) and 86.5% (2,566 characters) of them could be described. There are 399 characters left, but this includes 151 already existing named "Bushu". This result shows the good capability of the proposed scheme. A Formalism for Hierarchical Outline-Fonts Beat Stamm (Institute for Computer-Systems, Swiss Federal Institute of Technology, 8092 Zurich, Switzerland; Email: stamm@inf.ethz.ch) Abstract Simple font-rasterisation at low resolution is unlikely to preserve elementary properties of glyphs and characters due to inadequate control over quantization effects. Notably regularity on pixel-level is violated, and subtle nuances at high resolution are enlarged at low resolution, or tiny parts are dropped altogether. In this paper, we elaborate a concise formalism for defining outline-fonts, together with their regularity and similarity properties, and their behaviour under scaling. With this knowledge a simple font-scaling algorithm can consistently render stems, serifs and diacritics without further hints, instructions or grid-fitting. It does so by applying transformations to contours, glyphs or characters in pixel-space. Drop-outs and dynamic regularization are both taken care of by the contour's knots containing the intelligence to be scaled individually. The extensibility of this mechanism is asserted by object oriented implementation of the relevant parts. Advanced Typography Peter Karow (URW Unternehmensberatung, Harksheider Strabe 102, D-2000 Hamburg 65, Germany. Fax: +40 60-60-52-52) Abstract Recent research has brought us to computer aided typography. Especially, there are three interrelated innovations: 1. "Kerning on the Fly" calculates balanced kerning for words with a speed of 100 paris/sec. 2. The KQ-program fulfills "optical scaling" which calculates the various pointsizes by following the rules of the old punch cutters 3. The hz-program, finally, composes text without changing the inter- word-spacing following the rules of Johannes Gutenberg and his 42-line bible. Examples will be presented to prove the performance. Manufacturers of printing equipment should be concerned with these results and are called to improve its type-setting quality. The Design and Implementation of CCPE Jianping Fan (National Research Center for Intelligent Computing Systems, Institute of Computing Technology, Academia Sinica, P.O. Box 2704, Beijing, P.R. China) Abstract The CCPE (Chinese Character Pattern Editor) is an interactive Chinese Character (CC) design system of which focus on designing new typeface of CC fronts composed by 7-10 thousand CC patterns. In this paper, we discuss the general considerations or criterions in developing these kind of systems and their impacts in the design of the CCPE. The limited parametric graphical method and the character dynamic decomposition concept which be supported by CCPE are introduced in detail. The implementation of CCPE is presented in some detail. In the end of paper, some experiences gotten through using CCPE and future research work are highlighted. A Method of Binarization Based on Information of Character Structure Ning Sun (Computer Center, Tohoku Univ., 1-1, 2-chome, Aoba-ku, Sendai- shi 980, Japan. Fax: +81 22-263-9202), et al. Abstract Binarization of gray level image is the first process of character recognition and plays a crucial role in the recognition system. Until now, attempts of binarization are basically based on the methods using fixed threshold. The algorithm of using fixed threshold is simple and the effectiveness has been confirmed in respect to characters such as alphabets and simple Chinese characters, but it can be difficult to obtain a clear binary image from a complicated Chinese character especially where the character structure is proposed. The new method has two features, the first one is using a cubic B-spline surface in normalizing process, and the second one is the algorithm of using relative threshold based on local distribution of gray level image. As a result, the recognition system using the new method processed a first recognition rate of 99.71% in comparison with that of 99.32% using the typical binarization method. SGML-based Document Sharing System Daniel Oria (Engineering Research Lab., Matsushita Graphic Communication Systems Inc., 2-3-8 Shimomeguro, Meguro-ku, Tokyo 153, Japan Fax: +81 3-3495-4620), et al. Abstract We propose a SGML-based document sharing system which provides users with an integrated environment for electronic document processing and information sharing. The system, called SRIM (Strategic Reporting and Information Management), is basically designed for distributed use of PCs connected by PC-LAN to file and printer servers. The system can be composed of any other network configuration as well. The distinctive features of this system are creation and interchange of SGML documents. SGML offer the advantage of structure handling to process and share documents. The SRIM system contains the functionalities of editing and formatting SGML documents, documents registration, integration, and database management. ELIS (Electronic LIbrary System): A HyperBook System Mee-Young Sung (Computer Research Dept., Electronics and Telecommunications Research Institute), DaeDuk Science Town, Daejeon 305- 606, Korea; Email: mysung@kiet.etri.re.kr) Abstract ELIS (Electronic LIbrary System) is a HyperBook system designed for hypertexitization of conventional books and papers. It organizes information bits with links and provides a flexible multimedia presentation of them. ELIS supports four different types of links: associative link, detail link, note link, and external link. In addition, it is capable of such functions as authoring, browsing, searching, overview, backtracking, footprint/history, and bookmark. We defined a HyperBook description language, named ELIDL (Electronic LIbrary Description Language), to be used in generating ELIDF's (Electronic LIbrary Description Formats). Use of ELIDL makes creation and modification of HyperBooks simple because, with it, the main program need not be directly changed. A ELIDF is generated as an output from the authoring function which in turn becomes the input to the browsing function. ELIDF contains the structural informations of a HyperBook such as the types and positions of nodes, link types, and jump addresses. Technical Documentation Handling within a Computer Assisted Design Environment Francois Prunet (Dept. of Knowledge Acquisition and Representation, Laboratoire d'Informatique, de Robotique et de Microelectronique de Montpellier, Centre National de la Recherche Scientifique, U.A. 1480, University of Montpellier, 860 Rue de Saint Priest, 34090 Montpellier, France; Email: nanard@crim.fr or prunet@frmop53.bitnet), et al. Handling in real time the documentation associated to a design process is a very important challenge for high technologies industries. An integrated Computer Assisted Design environment should provide the designer with integrated tools for document production and management since documents are part of the design process: documents of any kind should be tightly linked to the objects they refer to. Structured hypertexts are a possible key to this challenge: they provide with flexible structure management. Typed links, typed nodes, typed anchors and dynamically specifiable relationships between types are powerful tools for representing knowledge about documents. The paper describes the architecture and the integrated CAD environment for Production Control Systems we have developed. It takes advantage of structured hypertext abstract machine both as a flexible repository for any information related to the design process and as a semantic link server which unifies the management of relationships between objects and their documentation. Multilingual Document Preparation Environment and Its Tools Zheng Min (Institute of Computer Science and Technology, Peking University, Beijing 100871, China; Email: zhm@compsci.liverpool.ac.uk.), et al. Abstract Establishing an environment instead of an isolated system is a new approach to the development of document preparation system. A multilingual document preparation environment -- WITS (Window-based Integrated Typesetting System) is described in this paper. It consists of a set of tools developed following predefined common data structures and communication protocols. These tools can work cooperatively with uniform operating style under a uniform environment. This paper gives a brief introduction to WITS environment, including an object-oriented platform and varieties of tools. Because WITS is a multilingual, especially multi-oriental language environment, discussion on multilingual processing is also included in this paper. On the Development of an Arabic/Latin Desktop Publisher M.N. Mikahail (The Center of Excellence, Computer Science Department, the American University in Cairo, Cairo, Egypt; Mikhail@auc.eg), et al. Abstract In spite of the important role Desktop Publishers play in our day- to-day business, there has been a lack of a good Arabic Desktop Publisher on the PC. In developing one, many major design issues had to be addressed. Manipulation of Arabic text in graphics mode with its unique features has been addressed. Creating a user-friendly lay-out utility, that can be dynamically changed, has been accomplished. Mixing Arabic, Latin, and bit-mapped graphics on the same document was also a challenging task. In order to minimize the secondary storage required to save the document, the layout is stored in a vector form. In addition, both Arabic and Latin text are stored as ASCII characters, and images have been compressed. Only when the document is to be viewed on the screen or sent to the printer, that the whole thing is converted to the graphics mode. Editing Facilities of Fine Multi-lingual Gray-scale Text Satoshi Abe (Faculty of Science, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113), et al. Abstract This paper discusses the implementation of text edit facilities with mixed-language gray-scale fonts. Text editing facilities are fundamental functions for human-computer interaction. However, even today's state- of-the-art display does not have sufficient resolution to display small Japanese characters that is essential for the business use. Gray-scale fonts can eliminate this problem. On the other hand, text editing of gray-scale fonts has several problems that are not found in the case of conventional binary fonts. The paper presents our implementation of text editing facilities for gray-scale fonts including the discussion on such problems. Audio Graphic Conference System Using Color LCD Tablet Kazuo Yoshimi (NTT Integrated Communications Systems Headquarters, 66-2 Horikawa-cho, Saiwai-ku, Kawasaki-shi, Kanagawa 210, Japan), et al. Abstract This paper describes an audio graphic conference system that transfers stored images, realtime voice and realtime handwriting between remote locations over an ISDN line. This system displays images concurrently at two remote locations for joint scrutiny by the parties to the conference. Either side can request an image, and it is displayed simultaneously on both side. Telewriting capability for over-writing is available. The main capabilities of the system are (1) realtime two-way voice transmission, (2) image transmission and display of the same images on LCD tablets on both sides, (3) realtime two-way telewriting and display on top of images, different colors are used to distinguish the notations of each side, (4) user-friendly human interface using a color LCD tablet. The displays support a full range of vivid colors. Moreover, the system has made much more user-friendly to operate because most functions can be controlled right on the tablet. The system should find extensive use in many new "tele-" domains. The NCS Colour System from Different Viewpoints Bjorn Kruse (Image Processing Group, Dept. of Electrical Engineering, Linkoping University, S-581 83 Linkoping, Sweden. Fax: +46 13-28-25-99) Abstract Colour systems can be classified with regard to their intended use. This is reflected in its representation of colour, that can be more or less adapted to the human colour vision. Some of the derivatives of the CIE-standard have been developed with the intent to adapt the representation to the visual properties of man. Being derivative of this system these representations are not as suitable as those developed with this intent in mind from the beginning. Such systems are usually referred to as colour appearance systems. It is required of a colour appearance system that the colour vector is represented in perception co- ordinates and that the scales are uniform in perception unit. This paper compares the natural colour system, NCS, with the CIELUV and OSA systems, of which the OSA is another colour appearance system. The comparison is made in graphical form to allow for a natural way to view colour populations in terms of scales and density with respect to the representation. --------------------------------END OF REPORT------------------------------