1、Reference number ISO/TR 14672:2000(E) ISO 2000 TECHNICAL REPORT ISO/TR 14672 First edition 2000-07-15 Graphic technology Statistics of the natural SCID images defined in ISO 12640 Technologie graphique Statistique des donnes dimages en couleur normales (SCID) dfinies dans lISO 12640ISO/TR 14672:2000
2、(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 installed on the computer performing the editing. In downloading this fi
3、le, 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 products used to create this PDF file can be found in the General Info r
4、elative 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 found, please inform the Central Secretariat at the address given below. IS
5、O 2000 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 in writing from either ISO at the address below or ISOs member body in the
6、 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.ch Web www.iso.ch Printed in Switzerland ii ISO 2000 All rights reservedISO ISO/TR 14672:2000(E) iii Contents Page Foreword iv Introduction v 1 Scope .1
7、2 Reference . 1 3 SCID image description .1 4 Colour distribution .3 4.1 Data value histograms and related things .3 4.2 Average colour values .5 4.3 4 x 4 covariance matrix 5 4.4 Three-dimensional volumes .6 5 Spatial frequency characteristics .7ISO/TR 14672:2000(E) ISO iv Foreword ISO (the Interna
8、tional Organization for Standardization) is a worldwide federation of national standards 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 be
9、en established has the right to be represented on that committee. International organizations, 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 stan
10、dardization. International Standards are drafted in accordance with the rules given in the ISO/IEC Directives, Part 3. 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 fo
11、r voting. Publication as an International Standard requires approval by at least 75 % of the member bodies casting a vote. In exceptional circumstances, when a technical committee has collected data of a different kind from that which is normally published as an International Standard (“state of the
12、 art”, for example), it may decide by a simple majority vote of its participating members to publish a Technical Report. A Technical Report is entirely informative in nature and does not have to be reviewed until the data it provides are considered to be no longer valid or useful. Attention is drawn
13、 to the possibility that some of the elements of this Technical Report may be the subject of patent rights. ISO shall not be held responsible for identifying any or all such patent rights. ISO/TR 14672 was prepared by Technical Committee ISO TC 130, Graphic technology, Working Group 2, Prepress data
14、 exchange.ISO ISO/TR 14672:2000(E) v Introduction International Standard 12640, Graphic technology Prepress digital data exchange CMYK standard colour image data (CMYK/SCID), provides the digital data for a set of natural and synthetic colour images. The natural images are intended for use in subjec
15、tively evaluating image quality as a function of image processing and/or output process. In addition, the synthetic images are provided to allow objective measurement of process control, tone reproduction, colour characterization, etc. In addition, these images may be used for the statistical evalua
16、tion of the change in image content as a function of image processing or other imaging steps. TC 130 Working Group 2 agreed to provide a set of reference statistics for these images as a baseline for the technical community wishing to use the images for such statistical evaluation. This statistical
17、data was prepared by the TC 130 Japanese National Committee, and their efforts have made this technical report possible.1 Graphic technology Statistics of the natural SCID images defined in ISO 12640 1 Scope This Technical Report provides the colour and spatial frequency distribution statistics asso
18、ciated with the digital image data of International Standard 12640, Graphic technology Prepress digital data exchange CMYK standard colour image data (CMYK/SCID). 2 Reference ISO 12640:1997, Graphic technology Prepress digital data exchange CMYK standard colour image data (CMYK/SCID). 3 SCID image d
19、escription ISO 12640 defines a set of natural and synthetic colour images, called SCID (Standard Colour Image Data), which can be used for evaluation of image processing algorithms or output devices, and also for colour characterization of output devices. These images are digital files, encoded in a
20、 CMYK format. They exist in two forms, known as the primary and alternate data sets, each of which has different resolutions and data ranges. The primary data set was created using the data encoding scheme typical of a Colour Electronic Prepress System (CEPS). The alternate data set was created from
21、 the primary data set by re- encoding the data in the scheme more typically used by the desktop publishing (DTP) prepress systems and defined in ISO 12639 as the P1 profile. With the primary set, tone value of 0% is encoded as data value of 28 and tone value of 100% is encoded as data value of 228;
22、the corresponding data values for the alternate set are 0 to 255. The primary set has an assumed resolution of 16 pixels/mm while the alternate set has an assumed resolution of 12 pixels/mm. The alternate set was computed from the primary set by cubic interpolation combined with a linear mapping of
23、the data values. The eight natural colour images of the primary set are shown as N1 to N8 in figure 1, the natural colour images of the alternate set are denoted as N1A to N8A. When output at the assumed resolutions of 16 pixels/mm (406.4 pixels/inch) and 12 pixels/mm (304.8 pixels/inch) the resulta
24、nt size is 160mm x 128mm. The primary set of natural images are 2560 x 2048 pixels and the alternate set are 1920 x 1536 pixels. Table 1 shows the characteristics and typical usage of the images. These eight images were selected so that users can subjectively evaluate several aspects of image qualit
25、y that are often reduced by image processing or as a result of output device characteristics. The image quality attributes in question are described in table 1, but may be summarized as follows: colour reproduction; sharpness; graininess. Although there is no established procedure for quantitative e
26、valuation of these subjective image quality attributes, by the use of a common set of images SCID makes it possible to conduct such subjective evaluation on a consistent basis. However, when evaluating the application of image processing algorithms to the images some quantitative analysis is feasibl
27、e and in this context various statistical characteristics calculated from the images may prove useful. Statistics on colour and spatial frequency distribution are such basic characteristics. TECHNICAL REPORT ISO ISO/TR 14672:2000(E)ISO/TR 14672:2000(E) ISO 2 N1 and N1A N2 and N2A N3 and N3A N4 and N
28、4A N5 and N5A N6 and N6A N7 and N7A N8 and N8AFigure 1 Reduced monochrome reproductions of the natural imagesISO ISO/TR 14672:2000(E) 3 Table 1 Natural images Name Aspect Characteristics N1, Portrait Portrait Used to evaluate the reproduction of human skin. N2, Cafeteria Portrait Image with complica
29、ted geometric shapes. Suitable for evaluating the result of image processing. N3, Fruit Basket Landscape Image of a basket and cloth used to evaluate the reproduction of brown colours and close texture. N4, Wine and Tableware Landscape Image of glassware and silverware used to evaluate the reproduct
30、ion characteristics of highlight tones and neutral colours. N5, Bicycle Portrait Image of a (penny-farthing) bicycle, resolution charts and other items containing high detail used to evaluate the sharpness of reproduction and the results of image processing. N6, Orchid Landscape Image of an orchid w
31、ith background vignettes used to evaluate reproduction of highlight and shadow vignettes. N7, Musicians Landscape Image of three girls with different skin characteristics and fine detail used to evaluate the reproduction of different skin tones and image detail. N8, Candle Landscape “Low-key” image
32、of a room scene containing miscellaneous objects used to evaluate dark colours, particularly browns and greens. These characteristics can be used for the comparison between the data of an original image and that of a transformed or degraded image passing through some image handling system such as an
33、 image transmission line, an imaging system, an image storage system or an image transform and processing system. It is also possible to compute those basic quantities from a set of images to certify validity of them if this is in doubt. Members of the technical committee ISO TC 130 evaluated the af
34、oresaid basic statistical quantities for the eight natural SCID images (for both the primary and alternate set). This technical report summarizes the quantitative results. 4 Colour distribution Statistics which represent image colour distribution are provided in this section. These were obtained by
35、computing single dimensional histograms, average colour values of each colour, the three-dimensional volume and covariance matrix and resultant orthogonal matrix and principal axes. Such data were calculated for each image. The histogram data is provided graphically (figures 5-12) and numerically (t
36、ables 5-12). The remaining statistics are provided in tables 13-20. It should be noted that each natural image is provided with a text insert such as ISO300 or ISO400 in the image. Pixels representing this text have a coded value of either 0 (white) or 255 (black). This text serves to distinguish be
37、tween the primary set and alternate set. It is not meaningful to include this text in the colour distribution calculation. Therefore, the calculation was carried out only for the image area, excluding the text. The position of the outer boundaries of the text is defined by a rectangle produced from
38、the coordinates of two of the corners as shown in figure 2. The position (in terms of number of pixels) of the text in each image is given in table 2 and table 3. This was the area excluded prior to calculation of the statistical data but extended by 4 pixels in each direction to minimize any effect
39、s arising from evaluation of the statistics after any image processing which requires calculation based on adjacent pixels. Thus the coordinates of the corners defining the rectangle enclosing the excluded area are A(X 1 -4,Y 1 -4) and B(X 2 +4,Y 2 +4). 4.1 Data value histograms and related things T
40、he histograms for the natural images are shown in figures 5-12. These show the frequency of occurrence of each of the dot percent equivalent values within the image; therefore, there are four histograms for each image, one for each of the colours cyan, magenta, yellow and black. Each of the figures
41、has two parts; the first shows the histograms for the primary set of images and the second part the histograms for the alternate set. The numeric data used to produce these figures are given in tables 5-12.ISO/TR 14672:2000(E) ISO 4 Figure 2 Definition of the coordinates of the text elementsTable 2
42、Position and area of the text for the primary set of natural images (16 pixels/mm )Image A(x 1 ,y 1 ) B(x 2 ,y 2 )N1 (1769,39) (2008,88)N2 (37,35) (276,84)N3 (38,36) (277,85)N4 (41,37) (280,86)N5 (37,34) (276,83)N6 (37,35) (276,84)N7 (2286,35) (2525,84)N8 (2070,193) (2309,242) Table 3 Position and a
43、rea of the text for the alternate set of natural images (12 pixels/mm )Image A(x 1 ,y 1 ) B(x 2 ,y 2 )N1A (1326,24) (1503,61)N2A (26,23) (203,60)N3A (25,22) (202,59)N4A (31,28) (208,65)N5A (25,23) (202,60)N6A (25,23) (202,60)N7A (1708,21) (1885,58)N8A (1554,145) (1731,182) It should be noted that al
44、though the primary set of images contains data in the ranges 0-27 and 229-255, these are not normally differentiated by the output device since the values less than or equal to 28 are all set to 0% dot and those greater than or equal to 228 are all set to 100% dot. In the case of the alternate set,
45、however, all values in the range may be differentiated. This means that when the alternate set was derived from the primary set all the values in the ranges 0-28 and 229-255 for the primary set were mapped to 0 and 255 respectively. Thus the histograms for the alternate set tend to show very high fr
46、equencies for these values when compared to the rest of the values within the image, or the values within the primary set. This phenomenon has a marked effect on some of the subsequent statistics, depending on the distribution of the out of range values in the primary images. In the case of the aver
47、age value, for example, it might be expected that the same (or very similar)ISO ISO/TR 14672:2000(E) 5 c j c i N , m j m i N , y j y i N , k j k i N M j (c im is the value for magenta; y is the value for yellow; k is the value for black; i is the index value of the pixel; and N is the number of pixe
48、ls in the image. The average values indicate the amount of each colour present in an image. Therefore, any tendency for a particular colour to be dominant in each image can be indicated, although not categorically stated, by the average colour values. For example, the high blue content of image N6 a
49、nd N6A (Orchid) is shown by the fact that the cyan and magenta colour values are higher than those of the yellow. However, similar values would have been obtained from an image containing large areas of cyan and magenta and no blue. Since no under colour removal (UCR) or gray component replacement (GCR) has been applied to these images black colour is found only in the shadow areas, and the average value is less than that of the coloured inks. However, the magnitude of the value provides some indication of whether an
