Proceedings of IAPR Workshop on MACHINE VISION APPLICATIONS

(IAPR MVA '90)

November 28-30, 1990, Kokubunji, Tokyo, Japan


Table of Contents

Front Messages in the Proceedings

Machine Vision

Recent Progress in Industrial Machine Vision (Invited) 1
Omid Mohtadi* and Jorge L. C. Sanz** *, *IBM Argentina (Argentina), **IBM Almaden Research Center, USA
Defect Detection, Classification and Quantification in Optical Fiber Connectors 15
Behzad Shahraray*, Arthur T. Schmidt* and J. Mark Palmquist, *AT&T Bell Laboratories, **AT&T Network Systems; USA
Precise Position Detection of a Mesh-Backed Printing Screen Based on the Smallest Enclosing Circle Detection 23
Takanori Ninomiya and Yasuo Nakagawa, Hitachi, Ltd., Japan

Industrial Applications

An Integrated Approach to Segmentation of Range Images of Industrial Parts 27
Arun K. Sood and Ezzet Al-Hujazi, George Mason University, USA
Shape and Strain Measurement of 3-D Object Using Fourier Transform Grid Method 31
Yoshiharu Morimoto, Li Tang and Yasuyuki Seguchi, Osaka University, Japan
New Techniques for 3-D Visualization and Object Classification Using Fan-Beam Sonars
Rajendar Bahl, Naval Postgraduate School, USA
Radar Image Processing for Locating Underground Linear Objects 35
Toru Kaneko, NTT, Japan
Interpretation of Industrial Scenes by Semantic Networks 39
Heinrich Niemann, Heinz Brünig, Richard Salzbrunn and Stefan Schröder, Universität Erlangen-Nürnbergt, FRG

Poster Session I

Method and Its Trial Instrumentation for High Speed and Continuous 3-D Measurement 43
Kazuo Araki*, Masanori Sato**, Takayuki Noda**, Yuji Chiba** and Masaru Shimizu**, *Chukyo University, **Nagoya Institute of Technology, Japan
Relaxation Based Modeling and Recognition of 3D Surfaces from Range Data 47
Chang Y. Choo*, Nasser M. Nasrabadi* and William I. Kwak**, *Worcester Polytechnic Institute, **Digital Equipment Corporation, USA
Recognition and Relocation of 3-Dimensional Objects Using Stereo Robot Vision System 51
Kyoung Mihn Do, Chee Woo Kang and Kwae Hi Lee, Sogang University, Korea
Shape Inspection of 3-D Objects Using Time-Coded Pattern Projection and Newly Developed Image Sensor Systems 55
Saburo Okada*, Tetsuhiro Sumimoto*, Hidekazu Miyauchi*, Masaaki Imade* and Hideki Yamamoto**, *Government Industrial Research Institute, Chugoku, **Okayama University, Japan
Image Sequence Analysis of Real World Human Body Motion 59
Xin Zhou, Qing Lu and Zhe Gu, Fudan University, PRC
Determining Surface Orientation by Uni-directional Photometric Flow Fields 63
Hidekazu Tanaka*, Yuji Iwahori** and Naohiro Ishii**, *Gifu Vocational Institute of Technology, **Nagoya Institute of Technology, Japan
A Change Detection Approach for Urban Changes from Remotely Sensed Data Using the Random Field Model *
Kiron K. Rao and R. Krishnan, Advanced Data Processing Research Institute, India
A Flexible Approach to the Construction of Expert Systems for the Application of Image Processing Software 67
Horst Bunke*, Felix Grimm* and Song Maoqiang**, *Universität Bern, **Beijing University of Posts and Telecommunications (PRC), Switzerland
Recognition of Partially Obscured Family of Objects Stretched along Axis of Symmetry 71
David T. Wang, Jaw-Chuan Ho, Jyh-Woei Chen, Shwu-Jin Lay and Peter A. Ng, New Jersey Institute of Technology, USA
Application of Image Reconstruction by Means of Chirp z-Transform 77
Kunio Takaya*, Tie-nan Ma*, Koichi Shimizu**, Masataka Kitama** and Tomohisa Mikami**, *University of Saskatchewan, **Hokkaido University (Japan), Canada
A New Method for Texture Segmentation *
Xining Zhang, Nanning Zheng and Haian Zhu, Xi'an Jiaotong University, PRC
A New Method for Recognizing and Locating Objects by Searching Longest Paths 81
G. Bertrand and F. Boussofiane, Groupe ESIEE, France
Finding Stomach in X-ray Film Using Computational Geometry 85
Kazuo Aisaka, Kiyoshi Arai and Akihide Hashizume, Hitachi, Ltd. , Japan
An Automatic Adjustment System for VCR Magnetic Heads on Cylinder Units 89
Hiroshi Tsukada, Tadanori Komatsu, Sumio Nagashima and Teruhisa Fukazawa, Toshiba Corporation, Japan
Targetting Number Plates Effectively Using Sparse/Full Templates and Coarse/Fine Template Matching 93
Harvey A. Cohen and Alan L. Harvey, La Trobe University, Australia
Plate Flatness Meter Based on Moire Fringes 97
Jiro Matsuo, Toshiyuki Matsumi and Kouichi Kitamura, Nippon Steel Corporation, Japan
High Level Loader and Recognizer of Electrical and Electronic Images Diagrams 101
Edison Oliveira de Jesus, Escola Federal de Engenharia de Itajubà, Brazil
Analysis of Dimension in Mechanical Engineering Drawings 105
Suzanne Collin and Dominique Colnet, CRIN - INRIA Lorraine, France
Improving the Performance of Multilayer Perceptron through Empirical Maximum Likelihood (EML) Learning Rule 109
S. Semnani and M. J. J. Holt, Loughborough University of Technology, UK
A Pattern Classifier Integrating Multilayer Perceptron and Error-Correcting Code 113
Haibo Li, Torbjörn Kronander and Ingemar Ingemarsson, Linköping University, Sweden
Optimal Input Selection of Neural Networks by Sensitivity Analysis and Its Application to Image Recognition 117
Hiroshi Takenaga, Shigeo Abe, Masao Takatoo, Masahiro Kayama, Tadaaki Kitamura and Yosiyuki Okuyama, Hitachi, Ltd. , Japan
2n-Tree Classifiers for Realtime Image Segmentation 121
Byron E. Dom and David Steele, IBM Almaden Research Center, USA
Associative Contour Processing 125
Avidan J. Akerib and Smil Ruhman, Weizmann Institute of Science, Israel

3-D Shapes

Multi-Precision Position Measuring Method with R-HPSD Scheme 129
Masanori Idesawa, Riken: The Institute of Physical and Chemical Research, Japan
Extracting Shape and Reflectance of Glossy Surfaces by Using 3D Photometric Sampling Method 133
Hideichi Sato, Shree K. Nayar and Katsushi Ikeuchi, Carnegie Mellon University, USA
A Method for the Synchronized Acquisition of Cylindrical Range and Color Data 137
Yasuhito Suenaga and Yasuhiko Watanabe, NTT, Japan
Stereo Reconstruction through Disparity Space 143
Richard Baldwin*, Katsuhiko Sakaue**, Hiromitsu Yamada** and Kazuhiko Yamamoto**, *Polytechnic of East London, **Electrotechnical Laboratory (Japan), UK
Generation of 3-D Models Based on Image Fusion of Range Data 147
Yasuo Sakaguchi, Hirokazu Kato, Kosuke Sato and Seiji Inokuchi, Osaka University, Japan

Map and Line Drawing

Techniques for Line Drawing Interpretation: An Overview (Invited) 151
Rangachar Kasturi*, Senthil Siva* and Lawrence O'Gorman**, *The Pennsylvania State University, **AT&T Bel! Laboratories, USA
Knowledge-Based Interpretation of Road Maps Based on Symmetrical Skeletons 161
Markus Ilg and Robert Ogniewicz, ETH-Zentrum, Switzerland
Raster to Vector Conversion in a Map Interpretation System 165
Paul C. K. Kwok* and T. John Turner**, *The University of Calgary, **DataSpan Technology, Inc. , Canada
Geographical Feature Analysis Using an Integrated Graphic Information Processing System 169
Kazuaki Iwamura, Yasuei Nomoto, Shigeru Kakumoto and Masakazu Ejiri, Hitachi, Ltd. , Japan
Randomized Hough Transform (RHT) in Engineering Drawing Vectorization System 173
Pekka Kultanen*, Erkki Oja* and Lei Xu**, *Lappeenranta University of Technology, **Peking University (PRC), Finland

Image Sequence

Movement Vector Detection with Reliability Indices 177
Naoya Ohta, NEC Corporation, Japan
The Measuring System of a Marathon Runner 181
Nobuyuki Sasaki* and Iwao Namikawa**, *Toshiba Corporation, **Kansai Telecasting Corp. , Japan
Recognition of Human Motion Based on Interpretation of 2D Pattern Deformation 185
Keiko Shiraishi, Mutsuhiro Terauchi and Kenji Onaga, Hiroshima University, Japan
Object Motion Identification for Object Recognition 189
Vito Cappellini, Alberto del Bimbo and Paolo Nesi, Università di Firenze, Italy
An Application of Optical Flow - Extraction of Facial Expression - 195
Kenji Mase, NTT, Japan

Poster Session II

Descriptive Ability of Drawing Image Understanding Framework Using State Transition Models 199
Shin'ichi Satoh and Masao Sakauchi, University of Tokyo, Japan
Interpretation of Mechanical Engineering Drawings for Paper-CAD Conversion 203
Pascal Vaxivière* and Karl Tombre**, *ESSTIN, **CRIN - INRIA Lorraine, France
Development of an Automatic Recognition System for Plant Diagrams 207
Takashi Futatsumata, Go Shichino, Jun'ichi Shibayama and Akira Maeda, Mitsubishi Electric@Corporation, Japan
An Automatic Recognition System of Assembly Drawings 211
Yong-Qing Cheng, Yu-Long Cao and Jing-Yu Yang, East China Institute of Technology, PRC
Character Font Coding and Processing by Polycurve Codes
Chang Y. Choo, Worcester Polytechnic Institute, USA
Recognition of Roads in an Urban Map by Using the Topological Road-Network 215
Takuya Hayakawa*, Toyohide Watanabe*, Yuuji Yoshida* and Kunio Kawaguchi**, *Nagoya University, **Sony Corporation, Japan
Automated Height Information Acquisition from Topographic Map 219
Sukit Viseshsin and Shunji Murai, University of Tokyo, Japan
An Interactive Colour Line Recognition System for Seismic Section Digitisation 223
Jonathan Shapiro and Jin Zhengping, The Turing Institute, UK
Two Systems for Converting Raster Data to Numerical Control Data 227
Robert H. Thibadeau* and Dale M. McNulty**, *Carnegie Mellon University, **University of California, Irvine, USA
FIHT2 Algorithm: A Fast Incremental Hough Transform 233
Hiroyasu Koshimizu* and Munetoshi Numada**, *Chukyo University, **Lossev Technology Corp. , Japan
Finding Line Segments by Surface Fitting to the Hough Transform 237
Wayne Niblack and Tai Truong, IBM Almaden Research Center, USA
Automatic Fire Detection by Changing Region Analysis Based on Geographic Information 241
Hideo Numagami and Akio Okazaki, Toshiba Corporation, Japan
Precision Measurement in Low Resolution Imagery 245
D. C. Douglas Hung* and O. R. Mitchell**, *New Jersey Institute of Technology, **University of Texas at Arlington, USA
An Adaptive Algorithm for Edge Detection with Subpixel Accuracy in Noisy Images 249
Jesse S. Jin, University of Otago, New Zealand
A Comparison of Fiducial Shapes for Machine Vision Registration 253
Lawrence O'Gorman, Alfred M. Bruckstein, Chinmoy B. Bose and Israel Amir, AT&T Bell Laboratories, USA
Texture Discrimination Using an Adaptive Multiresolution Network Filter 257
Hiroshi Sakou, Hitoshi Matsushima and Masakazu Ejiri, Hitachi, Ltd. , Japan
Identification of Core and Delta Points in Fingerprint Images 263
V. S. Srinivasan, CMC Ltd. , India
Face Recognition without Features 267
Matthew Turk and Alex Pentland, Massachusetts Institute of Technology, USA
Identifying Regions of Interest in Complex Imagery 271
Ronald P. Blanford and Mark R. Thomsen, TRW Inc. , USA
Region Tracking through Neural Classifier 275
Patrick Thévenoux* ** and Pascal Vincent*, *Aerospatiale DE/E/ETRI, **ETCA/CREA/SP, France
A Unified Approach to Tracking, Motion and Structure from Motion in a Long Monocular Sequence
Zhengyou Zhang and Olivier Faugeras, INRIA Sophia-Antipolis, France
3-D Shape Reconstruction from Camera Motion with Inexact Motion Parameters 279
Keisuke Kinoshita and Koichiro Deguchi, University of Tokyo, Japan
Rigid and Nonrigid Motion Analysis: Robust Recovery of 3-D Structure and Motion 283
Hiroyuki Morikawa and Hiroshi Harashima, University of Tokyo, Japan

New Frontier

Automatic Classification of Fingerprint Images 287
Babu M. Mehtre* and Anil K. Jain**, *CMC Ltd. , **Michigan State University (USA), India
Feature Extraction from Front and Side Views of Faces for 3D Facial Model Creation 291
Taka-aki Akimoto, Richard Wallace and Yasuhito Suenaga, NTT, Japan
Natural Surface Characterization by Multifractals 295
Fabio Arduini, Stefano Fioravanti and Daniele D. Giusto, University of Genoa, Italy
Automatic Digitizing of the Colour-Layer of Thematic Maps 299
Rune Espelid, Nils Arne Alvsvaag, Jan Eileng and Ivar Skauge, IBM Bergen Scientific Centre, Norway
A Real Time Simulated Human Vision System Using Connectionist Models Applied to Target Tracking 303
Hidenori Inouchi*, Niall McLloghlin* and David Vernon**, *Hitachi Dublin Laboratory, **Trinity College, Ireland

Architecture

Architectures for Multidimensional Low- and Intermediate Level Image Processing (Invited) 307
Pieter P. Jonker, Erwin R. Komen and Robert P. W. Duin, Delft University of Technology, The Netherlands
Multiresolution Pyramid Architectures for Real-Time Motion Analysis 317
Peter J. Burt, David Sarnoff Research Center, USA
SimPar - A Programmer Oriented Software Environment for the Pyramidal Vision Machine SPHINX 323
Edwige E. Pissaloux, Samir Bouaziz, Alain Mérigot and Francis Devos, Université Paris XI, France
Mapping an N * N Image on P Processors 327
Erwin R. Komen, Wouter B. Teeuw, Robert P. W. Duin and Pieter P. Jonker, Delft University of Technology, The Netherlands
Single-Chip High-Speed Computation of Optical Flow 331
Per-Erik Danielsson, Par Emanuelsson, Keping Chen, Per Ingelhag and Christer Svensson, Linkoping University, Sweden

Navigation and Robot Vision

Localization of Autonomous Mobile Robots by Monocular Vision in Modelled Sites. Application to a Nuclear Power Plant 337
M. Dhome, P. Reis, G. Rives and M. Richetin, Blaise Pascal University, France
Using and Generating Environment Models for Indoor Mobile Robots 343
Yuval Roth-Tadak and Terry Weymouth, The University of Michigan, USA
Recognition of Parametrised Models from 3D Data 347
Ian Reid, Oxford University, UK
A Visual Servoing Approach Applied to Robotic Tasks 351
Jean-Philippe Urban, Guy Motyl and Jean Gallice, Blaise Pascal University, France
Estimating the Pose and Motion of a Known Object for Real-Time Robotic Tracking 357
Olli Silvén, University of Oulu, Finland

Poster Session III

Camera Calibration by Computational Projective Geometry 363
Kenichi Kanatani and Yasuhiro Onodera, Gunma University, Japan
When Should We Consider Lens Distortion in Camera Calibration 367
Yi-Ping Hung and Sheng-Wen Shieh, Academia Sinica, ROC
3D Structure from Motion Using Homocentric Spherical Spatiotemporal Image Analysis 371
Takayuki Yasuno and Teruo Hamano, NTT, Japan
Edge Visibility Regions - A New Representation of the Environment of a Mobile Robot 375
Raj Talluri and J. K. Aggarwal, University of Texas at Austin, USA
A Study of Moving Object Detection from Unstable Motion Picture 381
Hirofumu Ishii, Toshiyuki Kawara, Isao Kawahara and Makoto Fujimoto, Matsushita Electric Industrial Co. , Ltd. , Japan
Map-Driven Image Interpretation by Associative Model Indexing 385
Gianluca Foresti, Vittorio Murino, Carlo S. Regazzoni and R. Zunino, University of Genoa, Italy
Video-Rate Image Processing System for an Autonomous Personal Vehicle System 389
Mayumi Ohzora*, Tohru Ozaki*, Shigeru Sasaki*, Masumi Yoshida* and Yoshitaka Hiratsuka**, *Fujitsu Laboratories Ltd. , **Fujitsu Limited, Japan
Real Time Algorithms for Obstacle Avoidance by Using Reprojection Transformation 393
Zhigang Zhu and Xueyin Lin, Tsinghua University, PRC
HERMIA: An Heterogeneous and Reconfigurable Machine for Image Analysis 397
Gaetano Gerardi, Franco Chiavetta, Vito Di Gesù and Domenico Tegolo, Università di Palermo, Italy
Advanced Vision Processor with an Overall Image Processing Unit and Multiple Local Image Processing Modules 401
Hiroaki Kubota, Kazuhiro Fukui, Minoru Ishikawa, Hiroshi Mizoguchi and Yoshinori Kuno, Toshiba Corporation, Japan
Component Labeling on SIMD-SPMD Architecture 405
P. Duclos*, G. Giraudon**, P. Kaplan*, M. Auguin* and F. Boeri*, *Universite de Nice, **INRIA Sophia-Antipolis, France
Processing Elastic Patterns the Massively Parallel Way 409
Francis Van Aeken and Yves Demazeau, Vrije Universiteit Brussel, Belgium
Application of Hopfield Network to Human-Guided Contour Extraction. 413
Makoto Kato, Toyohisa Morita, Shin'ichiro Miyaoka and Makoto Noumi, Hitachi, Ltd. , Japan
Feedforward Static Neural Networks for Image Compression *
Roman W. Swiniarski, Nancy Arnonne and Jeff Feynman, San Diego State University, USA
Using Geometric Model Guided Active Contour Method for Traffic Sign Detection 417
Long Quan* and Roger Mohr**, *CRIN - INRIA Lorraine, **LIFIA-IMAG, France
A Computer Vision System that Assembles Canonical Jigsaw Puzzles Using the Euclidean Skeleton and Isthmus Critical Points 421
Roger W. Webster*, Paul W. Ross*, Paul S. LaFollette** and Robert L. Stafford**, *Millersville University, **Temple University, USA
A Coarse to Fine Corner-Finding Method 427
Carlo Arcelli*, Andreas Held** and Keiichi Abe**, *Istituto di Cibernetica, C. N. R. (Italy), **Shizuoka University, Japan
Handwriting Model Adjustable to Writers 431
Nobuyuki Kita, Electrotechnical Laboratory, Japan
Off-line Shorthand Recognition System 437
R. Ethindra Kumar, A. Agarwal and Arun K. Pujari, University of Hyderabad, India
Word Based Recognition of Cursive Script 441
P. V. S. Rao, Tata Institute of Fundamental Research, India
A High Speed Word Matching Algorithm for Handwritten Chinese Character Recognition 445
Katsumi Marukawa, Masashi Koga, Yoshihiro Shima and Hiromichi Fujisawa, Hitachi, Ltd. , Japan
Decision Feedback Equalisation for Document Input 451
John Bernsen, Philips Research Laboratories, The Netherlands
'Symbolic Subtraction' of Fixed Formatted Graphics and Text from Filled In Forms 457
Gerd Maderlechner, Siemens AG, FRG
Recognition of Document Structure on the Basis of Spatial and Geometric Relationships between Document Items 461
Qin Luo, Toyohide Watanabe, Yuuji Yoshida and Yasuyoshi Inagaki, Nagoya University, Japan

Document Images

Knowledge-Based System for Structured Document Recognition 465
A. Belaïd*, J. J. Brault** and Y. Chenevoy*, *CRIN - INRIA Lorraine, **Ecole Polytechnique de Montréal (Canada), France
Model Based Understanding of Document Images 471
Koichi Kise*, Ken'ichi Momota**, Masaki Yamaoka**, Jun'ichi Sugiyama**, Noboru Babaguchi** and Yoshikazu Tezuka**, *University of Osaka Prefecture, **Osaka University, Japan
A Segmentation Free Approach to Symbol Extraction and Recognition from Image Document 475
Maurice Milgram, Mattieu Jobert and Bertrand Lamy, University of Pierre & Marie Curie, France
Document Recognition System with Layout Structure Generator 479
Yoshitake Tsuji, Hiroyuki Kami, Masaaki Mizuno, Toshiyuki Tanaka, Haruhiko Tanaka, Masao Iwashita and Tsutomu Temma, NEC Corporation, Japan
Anatomy of a Page Reader 483
Henry S. Baird, AT&T Bell Laboratories, USA