Handbook of Image Quality, Characterization and Prediction.

【实例简介】
Handbook of Image Quality, Characterization and Prediction. Brian W Keelan. Marcel Dekker, 2002 好不容易找到的
Some material described in this book may be protected by one or more U.S. and/or foreign patents, its description herein should not be cons(rued as an implied license to use suci patented inventions The authors and publisher have taken care in the preparation of this book, but make no expressed or implied warranty of any kind and assune neither responsibilify for errors or omissions nor liability for incidental or consequential damage arising from the use of the information contained herein ISBN:0-8247-07702 This book is printed on acid-free paper titers Marcel dekker、It 270 Madison Avenue. New york. NY 10016 tel:212-696-9000;fax:212685-4540 Eastern Hemisphere Distribution Marcel Dekker Ag Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland l:41-61-261-8482;fax:4l-6l-261-8896 World wide web http://www.dekker.com The publisher offers discounts on this book when ordered in bulk quantities. For more infor mation,write to Special Sales/ Professional Marketing at the headquarters address above Copyright 2002 by Marcel Dekker Inc. All Rights reserved Neither this book nor any part may be reproduced or transmitted in any form or by any by any inMformation storage and retrieval system, without micr ofilring, and recording, or means, electronic or mechanical, including photocopying rmission in writing from the Current printing(last 10987654 PRINTED IN THE UNITED STATES OF AMERICA TLFebOoK This volume is dedicated to my wife, Eileen, for her constant support and encouragement throughout its preparation, and to my colleagues Paul, Dick, Karin, Bob, Scott, and Jack, with whom I have spent so many stimulating hours pondering the mysteries of image quality. MARCEL DEKKER, INC. 270 Madison Avenue. New York. New York 10016 TLFebOoK 日 MARCEL DEKKER, INC. 270 Madison Avenue. New York. New York 10016 TLFebOoK Preface The naturc and scope of imaging are undergoing dramatic change as it enters the digital era. Portions of the formerly distinct photographic, electronic, software, television, computer, and printing industries are converging into a more generic imaging industry. The ways in which images are used are increasing in number and diversity, and the flexibility associated with digital imaging is leading to an increasingly complex field of opportunity. The rapid product cycle time of the electronics industry sets the standard for the new imaging industry, leading to an urgent need to streamline strategic, design, and development processes. In this more horizontal industry, the ability to effectively exchange specifications and evaluations based upon a common framework will become critical to the success of supplier-manufacturer and partnering relationships Each of these industry trends is leading to an increasingly acute need for methods of quantifying, communicating, and predicting perceived image quality. Consider the following two cases that exemplify these needs. 1. The development of the Advanced Photo System(APS)was carried out by a consortium of five companies. Prediction of image quality through computer modeling was used to select the optimal format size to meet the image quality aims while enabling such new features as smaller cameras and panoramic prints. It was quickly recognized that the modeling could simulate the results of the truly definitive experiments that engineers and analysts would like to run but which would take far ≌ more time than was available. It was also soon appreciated that simplified experiments could yield mislcading rcsults, which, in fact, could be predicted from and explained by image quality modeling , ⅠN 270 Madison Avenue, New York. New York 10016 TLFebOoK Preface Subsequent to the critical step of format standardization, computer predictions were used for many purposes during component design, including: determining optimal film speed, specifying camera and printer lens and film performance requirements in terms of modulation transfer functions, setting tolerances for positioning of optical assembly subcomponents, establishing aims for film noise and identifying the most valuable information to record on the magnetic layer for improved photofinishing. Trade trials prior to system introduction proved that the image quality distributions predicted by the modeling closely matched thosc produccd in the marketplace 2. A new charge-coupled device(CCD)electronic sensor architecture was proposed for professional digital still camera applications. This proposal was demonstrably advantageous to sensor manufacture. Based on engineering rules of thumb, there was some concern that adoption of the architecture might place greater demands on other system components, some of which were manufactured by other companies Construction of prototype devices and subsequent testing and analysis would have rcquired cxccssivc cxpcnsc and time. Image quality modeling was used to predict the impact that the change in the sensor architecture would have on requirements for each of the other system components if the image quality were to remain unaffected. The revised tolerances and performance specifications were compared to achievable As result, plans for new manufacturing lines were canceled, and the detrimental effects arising from a complex set of system interactions re avoide These ify some of the applications of image quality modeling lechniques and suggest why the industry trends mentioned earlier are making such capabilities ever more critical. In general. the benefits of computer modeling are at least threefold 1. cycle time compression and cost savings through reduced prototyping and experimentation 2. identification of unexpected solutions that might be missed by empirical testing over a restricted range; and 3. education and training of practitioners through virtual experimentation ≌ At Eastman Kodak Company, prediction of image quality through computer modeling has proved to be of great value in all three regards and has been ⅠN 270 Madison Avenue, New York. New York 10016 TLFebOoK Preface v regularly used in formulating business strategies, guiding design decisions, establishing product aims, budgeting system tolerances, supporting advertising claims, and benchmarking competitors' offerings Despite such local successes, it is widely assumed that image quality, being a subjective attribute, is not amenable to quantitative analysis. This misconception is difficult to overcome because of several factors the infrequency of coverage of pertinent topics, such as psychometrics, in academic curricula 2. the absence of a published, integrated approach to image quality characterization and prediction; and 3. the scarcity of non-proprietary examples of image quality modeling that could be shared among the industrial and academic communities The present volume addresses these issues through a review of needed background material, a description of an integrated and comprehensive approach to image quality modeling, and the provision of a number of examples of applications. This book is intended particularly for image scientists and product engineers, but portions of it should prove useful to individuals involved in project management, manufacturing quality control, marketing, business research, systems performance anal ysis, human factors and usability assessment trade journal evaluationS, and standards definition. It is hoped that this publication will focus new attention on, and stimulate further advances in, the fascinating field of image quality. Brian w, eelan ≌ z哥a MARCEL DI IN 270 Madison Avenue. New York. New York 10016 TLFebOoK Acknowledgments The majority of the results described in this book are drawn from investigations carried out by karin Topfer, Paul J. Kane, Robert E. Cookingham, Richard B Wheeler. John E. Kaufman Scott F O Dell and the author with assistance from Andrew D. Thompson, Donna L Hofstra, James L, Miller, Stacey L. Mayo, and Sharon M. Skelly. The first four named individuals each co-authored three chapters in this volume and also assisted in its preparation in many additional respects This manuscript benefited from the constructive comments of Paul w. Jones, Scott F o'Dell. Katherine S. Marsh. Eileen L. Keelan, John V. Nelson Edward J. Giorgianni, David M. woods, and the chapter co-authors. R. Brian Porter and J. Monty Wright provided valuable assistance regarding electronic organization and page layout. Margaret L. Mauer consulted in the preparation of the index and Richard H. Repka wrote software for organizing the electronic hics file without the advice, encouragement and support offered by Brian J. Thompson, John V. Nelson. and James C. Weaver, it is doubtful whether this work could have been brought to fruition. Finally, I would like to thank my production editor eric F. stannard. and the other staff of marcel dekker Inc who contributed to the production of this volume Note: Figures14.2,14.3,22.2,26.1,26.8,30.1,30.4,and31.2 have been reprinted from the proceedings of IS&Ts PICS 2000 conference(Portland ≌ Oregon)with the permission of the Society for Imaging Science and Technology (Springfield, Virginia) z哥a MARCEL DEKKER, INC. 270 Madison Avenue. New York. New York 10016 TLFebOoK Introduction To crcatc a computer modcl capable of predicting the image quality that would be produced by a hypothetical imaging system, researchers at Eastman Kodak Company have taken the following steps 1. establishment of a numerical scale of image quality that is anchored to a set of physical standards (images) and is calibrated in perceptually useful terms that facilitate its interpretation just noticeable differences) 2. development of a psychometric measurement technique efficientl yielding reproducible results that are calibrated in terms of the standard le of image quality from Step #1 3. elucidation of a theory for the prediction of the overall(multivariate quality of an image from a knowledge of its individual quality attribute levels(e.g, sharpness, graininess, etc. 4. Investigation of a selected set of subjective image quality attributes(as in Step #3 )using the psychometric technique from Step #2, leading to the dcfinition of objective metrics(e, g, granularity) bearing a known relationship to calibrated assessments of the subjective attributes 5. implementation of propagation models(e.g, linear systems theory) that, from key properties of system components, predict the corresponding properties of final images, in support of computation of the objective metrics from Step #4 日 ARCEL DEKKER. INC. 270 Madison Avenue. New York. New York 10016 TLFebOoK 【实例截图】
【核心代码】