1、 Reference number ISO 20462-1:2005(E) ISO 2005INTERNATIONAL STANDARD ISO 20462-1 First edition 2005-11-01 Photography Psychophysical experimental methods for estimating image quality Part 1: Overview of psychophysical elements Photographie Mthodes psychophysiques exprimentales pour estimer la qualit
2、 dimage Partie 1: Aperu gnral des lments psychophysiques ISO 20462-1:2005(E) PDF disclaimer This PDF file may contain embedded typefaces. In accordance with Adobes licensing policy, this file may be printed or viewed but shall not be edited unless the typefaces which are embedded are licensed to and
3、 installed on the computer performing the editing. In downloading this file, parties accept therein the responsibility of not infringing Adobes licensing policy. The ISO Central Secretariat accepts no liability in this area. Adobe is a trademark of Adobe Systems Incorporated. Details of the software
4、 products used to create this PDF file can be found in the General Info relative to the file; the PDF-creation parameters were optimized for printing. Every care has been taken to ensure that the file is suitable for use by ISO member bodies. In the unlikely event that a problem relating to it is fo
5、und, please inform the Central Secretariat at the address given below. ISO 2005 All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying and microfilm, without permission i
6、n writing from either ISO at the address below or ISOs member body in the country of the requester. ISO copyright office Case postale 56 CH-1211 Geneva 20 Tel. + 41 22 749 01 11 Fax + 41 22 749 09 47 E-mail copyrightiso.org Web www.iso.org Published in Switzerland ii ISO 2005 All rights reservedISO
7、20462-1:2005(E) ISO 2005 All rights reserved iii Contents Page Foreword iv Introduction v 1 Scope. 1 2 Normative references. 1 3 Terms and definitions. 1 4 Specification of the experimental conditions and results 5 4.1 Observer characteristics 5 4.2 Stimulus properties 6 4.3 Instructions to the obse
8、rver 6 4.4 Viewing conditions . 7 4.5 Experimental duration 7 4.6 Results. 8 4.7 Summary of reported quantities . 8 Annex A (informative) Selection of an appropriate psychophysical method 9 Annex B (informative) Stimulus differences, paired comparison proportions, and JNDs . 11 Annex C (informative)
9、 Example of a report of a psychophysical experiment 13 Annex D (informative) Comparison of selected psychometric methods 15 Bibliography . 17 ISO 20462-1:2005(E) iv ISO 2005 All rights reservedForeword ISO (the International Organization for Standardization) is a worldwide federation of national sta
10、ndards bodies (ISO member bodies). The work of preparing International Standards is normally carried out through ISO technical committees. Each member body interested in a subject for which a technical committee has been established has the right to be represented on that committee. International or
11、ganizations, governmental and non-governmental, in liaison with ISO, also take part in the work. ISO collaborates closely with the International Electrotechnical Commission (IEC) on all matters of electrotechnical standardization. International Standards are drafted in accordance with the rules give
12、n in the ISO/IEC Directives, Part 2. The main task of technical committees is to prepare International Standards. Draft International Standards adopted by the technical committees are circulated to the member bodies for voting. Publication as an International Standard requires approval by at least 7
13、5 % of the member bodies casting a vote. Attention is drawn to the possibility that some of the elements of this document may be the subject of patent rights. ISO shall not be held responsible for identifying any or all such patent rights. ISO 20462-1 was prepared by Technical Committee ISO/TC 42, P
14、hotography. ISO 20462 consists of the following parts, under the general title Photography Psychophysical experimental methods for estimating image quality: Part 1: Overview of psychophysical elements Part 2: Triplet comparison method Part 3: Quality ruler method ISO 20462-1:2005(E) ISO 2005 All rig
15、hts reserved v Introduction There are many circumstances under which it is desirable to quantify image quality in a standardized fashion that facilitates interpretation of results within a given experiment and/or comparison of results between different experiments. Such information can be of value i
16、n assessing the performance of different capture or display devices, image processing algorithms, etc. under various conditions. There are a number of psychometric methods described in the literature, such as paired comparison, rank ordering, categorical sort, and magnitude estimation, which might b
17、e considered as candidates for experimentally measuring image quality. Several textbooks 1 3 4 5 9 12have reviewed these and other methods and have discussed associated data reduction techniques, which usually are based upon the approach of Thurstone 11or analogous reasoning. However, the choice of
18、the best method for a particular application may be difficult to make, and interpretation of the rating scales produced by the numerical analyses is frequently ambiguous. Furthermore, none of the commonly used techniques provides an efficient mechanism for calibration of the results against a standa
19、rdised numerical scale or associated physical references, which is desirable when results of different experiments are to be compared or integrated. The value of new calibrated psychometric methods in developing comprehensive models of imaging system quality has been demonstrated in a recent work 6t
20、hat contains more detailed discussions of many of the informative topics superficially considered herein. The three parts of ISO 20462 address the need for documented means of determining image quality in a calibrated fashion. Part 1 provides an overview of practical psychophysics; specific experime
21、ntal methods and associated data reduction techniques are described in Part 2 (triplet comparison method 8 10 ) and Part 3 (quality ruler method 6 ). Informative Annex A aids in identifying the better choice between the two alternative methods of Parts 2 to 3, which are complementary and together ar
22、e sufficient to span a wide range of applications. It is the intent of these methods to produce results that are not merely directional in nature, but are expressed in terms of relative or fixed scales that are calibrated in just noticeable differences (JNDs), so that the significance of experimenta
23、lly measured stimulus differences is readily ascertained. INTERNATIONAL STANDARD ISO 20462-1:2005(E) ISO 2005 All rights reserved 1 Photography Psychophysical experimental methods for estimating image quality Part 1: Overview of psychophysical elements 1 Scope This part of ISO 20462 is part of a mul
24、tiple-part standard pertaining to the subjective evaluation of pictorial still image quality. This part of ISO 20462 a) defines the units by which image quality is quantified (just noticeable differences, or JNDs); b) describes the influence of stimulus properties, observer characteristics, and task
25、 instructions on results obtained from rating experiments; c) provides a flow chart for choosing the preferred psychophysical method for determining image quality from among those defined in subsequent parts of ISO 20462. 2 Normative references The following referenced documents are indispensable fo
26、r the application of this document. For dated references, only the edition cited applies. For undated references, the latest edition of the referenced document (including any amendments) applies. ISO 3664, Viewing conditions Graphic technology and photography 3 Terms and definitions For the purpose
27、of this document, the following terms and definitions apply: 3.1 artefactual attribute attribute of image quality that, when evident in an image, nearly always leads to a loss of overall image quality EXAMPLE Examples of artefactual attributes include noise and aliasing. NOTE The commonly used terms
28、 defect and impairment are similar in meaning. 3.2 attribute aspect, dimension, or component of overall image quality cf. artefactual attribute (3.1) and preferential attribute (3.12) ISO 20462-1:2005(E) 2 ISO 2005 All rights reservedEXAMPLE Examples of image quality attributes include image structu
29、re properties such as sharpness and noise; colour and tone reproduction properties such as contrast, colour balance, and relative colourfulness; and digital artefacts such as aliasing, contouring, and compression defects. 3.3 attribute just noticeable difference attribute JND measure of the detectab
30、ility of appearance variations, corresponding to a stimulus difference that leads to a 75:25 proportion of responses in a paired comparison task in which univariate stimuli pairs are assessed in terms of a single attribute identified in the instructions cf. quality JND (3.14) NOTE 1 As an example, a
31、 paired comparison identifying the sharper of two stimuli that differ only in their generating system modulation transfer function (MTF), would yield results in terms of sharpness attribute JNDs. If the MTF curves differed monotonically and did not cross, the outcome of the paired comparison would d
32、epend primarily upon the observers ability to detect changes in the appearance of the stimuli as a function of MTF variations, with little or no value judgement required of the observers. The relationship between paired comparison proportions and stimulus differences is discussed in greater detail i
33、n Annex B. NOTE 2 If observers are instead asked to choose which of a pair of stimuli is higher in overall image quality, and if the stimuli in aggregate are multivariate, such that the observer should make value judgements of the importance of a number of attributes, rather than focussing on one as
34、pect of image appearance, it is observed experimentally that larger objective stimulus differences (for example, MTF changes) are required to obtain a 75:25 proportion of responses, which in this case corresponds to a quality JND. NOTE 3 A JND is a statistical quantity, derived from a number of obse
35、rvations. An observer assessing a single pair of images differing by one attribute JND is unlikely to be confident that he or she has detected the sample difference. A stimulus difference of approximately three JNDs is usually needed for an observer of average sensitivity to feel reasonably certain
36、of his or her assessment. 3.4 categorical sort method psychophysical method involving the classification of a stimulus into one of several ordered categories, at least some of which are identified by adjectives or phrases that describe different levels of image quality or attributes thereof NOTE The
37、 application of adjectival descriptors is strongly affected by the range of stimuli presented, so that it is difficult to compare the results of one categorical sort experiment to another. Range effects and the coarse quantization of categorical sort experiments also hinder conversion of the respons
38、es to JND units. Given these limitations, it is not possible to unambiguously map adjectival descriptors to JND units, but it is worth noting that in some experiments where a broad range of stimuli have been presented, the categories excellent, very good, good, fair, poor, and not worth keeping have
39、 been found to provide very roughly comparable intervals that average about six quality JNDs in width. 3.5 image quality impression of the overall merit or excellence of an image, as perceived by an observer neither associated with the act of photography, nor closely involved with the subject matter
40、 depicted NOTE The purpose of defining image quality in terms of third-party (uninvolved) observers is to eliminate sources of variability that arise from more idiosyncratic aspects of image perception and pertain to attributes outside the control of imaging system designers. 3.6 instructions set of
41、 directions given to the observer for performing the psychophysical evaluation task 3.7 just noticeable difference JND stimulus difference that leads to a 75:25 proportion of responses in a paired comparison task cf. attribute JND (3.3) and quality JND (3.14) ISO 20462-1:2005(E) ISO 2005 All rights
42、reserved 3 3.8 magnitude estimation method psychophysical method involving the assignment of a numerical value to each test stimulus that is proportional to image quality; typically, a reference stimulus with an assigned numerical value is present to anchor the rating scale NOTE The numerical scale
43、resulting from a magnitude estimation experiment is usually assumed to constitute a ratio scale, which, ideally, is a scale in which a constant percentage change in value corresponds with one JND. In practice, modest deviations from this behaviour occur, complicating the transformation of the rating
44、 scale into units of JNDs without inclusion of unidentified reference stimuli (having known quality) among the test stimuli. 3.9 multivariate describing a series of test or reference stimuli that vary in multiple attributes of image quality 3.10 observer individual performing the subjective evaluati
45、on task in a psychophysical method 3.11 paired comparison method psychophysical method involving the choice of which of two simultaneously presented stimuli exhibits greater or lesser image quality or an attribute thereof, in accordance with a set of instructions given to the observer NOTE Two limit
46、ations of the paired comparison method are as follows. a) If all possible stimulus comparisons are done, as is usually the case, a large number of assessments are required for even modest numbers of experimental stimulus levels if N levels are to be studied, N(N 1)/2 paired comparisons are needed. b
47、) If a stimulus difference exceeds approximately 1,5 JNDs, the magnitude of the stimulus difference cannot be directly estimated reliably because the response saturates as the proportions approach unanimity. However, if a series of stimuli having no large gaps are assessed, the differences between m
48、ore widely separated stimuli may be deduced indirectly by summing smaller, reliably determined (unsaturated) stimulus differences. The standard methods for transformation of paired comparison data to an interval scale (a scale linearly related to JNDs) perform statistically optimized procedures for
49、inferring the stimulus differences, but they may yield unreliable results when saturated responses are included in the analysis. 3.12 preferential attribute attribute of image quality that is invariably evident in an image, and for which the preferred degree is a matter of opinion, depending upon both the observer and the image content EXAMPLE Examples of preferential image quality attributes include colour and tone reproduction properties such as contrast and relative colourfulness. Because the perc
