EP1963103A1

EP1963103A1 - Methods for evaluating and improving print quality

Info

Publication number: EP1963103A1
Application number: EP06830946A
Authority: EP
Inventors: Jouni Marttila; Esa Torniainen; Pekka Laihanen; Karoliina SYRJÄNEN
Original assignee: M Real Oyj
Current assignee: Metsa Board Oyj
Priority date: 2005-12-21
Filing date: 2006-12-21
Publication date: 2008-09-03
Also published as: RU2008128476A; WO2007071826A1; FI20051312A0; FI20051312L

Abstract

The invention relates to a method for determining the quality of print quality. In the method, a first digital image, which represents the intended print quality in the graphical product, and a second digital image, which represents the achieved or planned print quality of the graphical product, are taken. After this, a variable proportional to the magnitude of the difference in the colour values of at least some image elements corresponding to each other in the image plane of the first and second digital images, is calculated, and on the basis of the variable, the similarity of the colour worlds of the first and second images is estimated. The invention permits the objective comparison of images, particularly in the development of print quality and in claims situations.

Description

Methods for Evaluating and Improving Print Quality

The present invention relates to a method for evaluating print quality. In particular, the method relates to a method, by means of which the colour reproduction of a so-called proof can be evaluated relative to the intended print quality, or by means of which the colour reproduction of a finished printed product can be evaluated relative to a proof.

Proofs play a significant role in the work processes of many printed products (e.g., magazines, printed advertising and illustrated catalogues and books). When producing printed material, a proof is used to predict what the printed product will look like. A proof can also be used to specify the quality of the printed product by instructing the printer to print according to a supplied proof (a so-called contract proof). This presupposes that the proof has been produced according to specific criteria, the most important of which in terms of colour reproduction are the colour-reproduction capability of the proofing device, as well as the so-called simulation profile (the 'print profile' of the proof) used to produce the proof. The purpose of the simulation profile is to achieve the same colour values in the proof as when printing.

The print profile (ICC profile) is a computer file, in which is recorded the connection between the CMYK Cyan, Magenta, Yellow, Black) values used in printing and the colour values (Lab values) measured from the printed field. Various computer programs use ICC profiles, for example, for converting from an RGB (Red, Green, Blue) space (the colour formation used by computer displays, televisions, and digital cameras, for example) to a CMYK space (the colour formation used by printing presses and computer printers). Like the RGB space, the Lab space is a three-dimensional colour model, which best depicts the colour world seen by the human eye.

Nowadays, proofs are also often used for examining the quality of a print profile, due to the fact that development work of this kind is considerably cheaper than making the same tests on a real printing press. The quality of a print profile can be characterized by means of two criteria. First of all, how well does the print profile correspond to the printing conditions? This also indicates the success of the test printing and the stability of the printing process. Secondly, how linearly does the print profile perform colour conversions? This can be examined, for example, by monitoring how even a colour scale, for example, from white to black, or from red to green, the print profile is able to produce.

Publication 2005/014673 discloses one method for correcting the colours of an output image to correspond to a printed image. In the method a comparison is made, on the one hand between the YMCK values used to make the output image and the colour values measured from the image and, on the other hand, between the YMCK values used to make the printed image and the colour values measured from the image. The method requires complete familiarity with the profiles used to produce both images.

A problem in comparing a proof and the printed product, or a proof and the original is that the comparison is very often based on a subjective evaluation of quality, which is affected by the professional skill and preferences of the person making the comparison and also to a great extent by the lighting conditions. Lighting conditions are critical particularly when comparing a proof made using inkjet technology with a printed product made using offset technology, due to the different pigments used by the technologies. Professional printers generally used a colour temperature of 5000 K when examining products. In many cases however, proofs are viewed in mixed light composed of, for example, the light of fluorescent light, halogen light, and daylight. However, even selected viewing conditions do not imply objective quality evaluation, as the value depends greatly on the observer.

Particularly in the case of claims and quality development, it would, however, be important to obtain objective information for the comparison of the proof and the printed product. In addition, it is not always known what print profiles were used to produce the proof or the printed product, so that their similarity must be evaluated without this information.

Publication US 2001/0017719 discloses one method for the subjective colour correction of printouts. In it, several images with modulated colours are output, from which the operator selects the image with a colour world closest to a reference image.

The invention is intended to substantially reduce the effect of subjective and ambient factors in the aforementioned proof comparison problem.

The invention is based on the idea that a proof in a digital form is compared with a reference image in a digital form, which represents the achieved or planned print quality of a graphical product, in such a way that on the basis of the proof and the colour values of the reference image, a variable is calculated, typically a difference or ratio, proportional to the difference in magnitude of the colour values of at least some of the image elements corresponding to each other in the image plane of the proof and the reference image. The same can be repeated for many of all of the image elements that correspond to each other, when a third digital image (a so-called colour-difference image) will be created. On the basis of the comparison variable or variables, the similarity of the colour worlds (colour reproduction, contrast, etc.) of the first and second images can be estimated more objectively.

The proof or reference images, generally both, are typically initially in a paper form. Thus they must be digitalized, for example using a scanner or camera prior to the creation of the colour-difference image. However, it should be noted that at no stage of the method does the proof and/or reference image necessarily have to be in paper form, or generally exist other than digitally. For example, by using the aforementioned ICC profiles, it is possible to simulate both images. The image original can be converted, for example, using two different print profiles or other image parameters, such as colour-temperature adjustments, into two images differing slightly from each other in their colour worlds, and with the aid of the present method an examination can be made as to whether the difference in colour world is significant in some specific application. An example is an application, in which the effect of different lighting conditions on the appearance of a printed product is investigated. The colour-difference image expresses objectively the parts of the image, in which a change in illumination may possibly cause visually discernible changes.

On the basis of what is stated above, it is obvious that, always depending on what the proof is compared with, it can represent the intended print quality, or the 'achieved' print quality (i.e. the print quality, which in an ideal case, i.e. a case in which the proofing device and the printing press would make an identical print quality, the printing press would produce). The reference image then represents correspondingly the achieved print quality, or the intended print quality (the image original).

The reference image is usually a second proof, image original, or finished printed product. Due to the many different purposes of the present method, the comparable images can be referred to hereinafter by the terms 'first image' and 'second image'. The colour-difference image can be created in several different ways and is generally referred to by the term 'third image'. In many of the later examples, the terms 'proof, 'printed product', and 'image original' will be used for reasons of clarity, but one skilled in the art will understand that the same principles can also be applied to other corresponding comparison situations involving two images.

More specifically, the method, according to the said invention, for evaluating print quality is characterized by what is stated in the characterizing portion of Claim 1.

More specifically, the method, according to the said invention, for improving print quality is characterized by what is stated in the characterizing portion of Claim 20.

The uses according to the invention are characterized by what is stated in Claims 21 and

22.

Considerable advantages are gained with the aid of the invention. It provides a new and more objective way to evaluate the quality of a printing or proofing process. It is well known that there are large differences in the abilities and ways of different people to see images, particularly colours and colour combinations. The reference variable produced by the method is not, however, dependent of the subjective characteristics of the evaluator's eyes or brain. Thus the evaluation of a printing or proofing process can be automated and its repeatability improved. The invention also provides a new way to produce objective criteria, which can be used to define a print quality being aimed at, as well as a new way of objectively evaluating the quality of the print quality of a printed product, for example, in cases of claims.

If a third image is created on the basis of the images being compared, it will be possible to see directly from this in which geometric point in the image there is a colour difference in the images being compared, and how important this difference is objectively. After this, the subjective importance of the colour difference can be further evaluated, either by calculation or otherwise, by using subjective parameters. By means of these parameters, it possible to take into account, for example, the way that the eyes and brain perceive colours and colour scales 'non-linearly' (the same objective colour difference in one specific point in a continuous tone scale is not as significant as the same colour difference at another point).

It should be noted that the invention is suitable for use with both real images (e.g., photographs, drawings, or products containing them) and with artificial test images (colour scales, colour segments, etc.), not does it as such require familiarity with the variables used in producing the images for comparison.

In the following, various embodiments of the invention are examined with reference to the accompanying drawings. In the drawings

Figure 1 illustrates different colour spaces, Figures 2a - 2c show two images differing slightly in colour world and correspondingly an image of colours calculated on their basis, and

Figures 3a - 3c show colour scales differing from each other and colour-difference images calculated from them.

The variable proportional to the difference in magnitude of the colour values of the corresponding image elements of the first and second image generally refers to the difference of the colour values of the image elements, or to a variable proportional to this, or to the ratio of the colour values of the image elements, or to a variable proportional to this. The colour difference is depicted with the aid of the parameter Delta E.

According to a preferred embodiment, the method according to the invention, at least the calculation of the variable or variables proportional to the difference in magnitude of the colour values, is implemented by means of software run in a computer. In the description below, an embodiment, in which a colour-difference image is created on the basis of the images being compared, is often used as an example. On the basis of the above description it is, however, obvious that the method does not require the entire image to be calculated, but instead a serviceable reference variable can also be calculated on the basis of individual image elements. The number of degrees of freedom of the colour values of the colour-difference image can be less that those of the original images. This means that the value of a single element of the colour-difference image can correspond to several directly calculated colour-difference values. Of course, on the basis of a directly calculated colour-difference image, it is also possible to calculate a new colour-difference image, the value space of which is reduced. Such an image will be particularly illustrative. One example of such an application is shown in Figures 2a - 2c. Figures 2a and 2b show a proof and a reference image. In Figure 2c is a colour-difference image calculated on the difference of the colour differences of Figures 2a and 2b, which reveals the differences in tone of the images. In the colour- difference image, the colour difference of the red (square) areas is greater than 5, that of the green (circle) area is less than 2, and that of the grey (triangle) areas is between those.

A colour-difference image can be created for all the channels of the image, i.e. for different colours (RGB, CMYK), or for the other dimensions of the colour space used (HST, Lab), either simultaneously or separately. In the first case, the values of colour-difference image represent the summed difference over the different channels. Naturally, in the case of grey- tone images it is possible to use only a single channel.

The image elements used in calculation can be, for example, individual pixels (pixel element), image elements (convolution element) obtained as a result of convolution (spatially filtered), or, for example, areas (subject areas) separated from the image by pattern-recognition means. Some mathematical operation, for example obtaining a mean value) can thus be performed on the colour values of several pixels, over a specific set before creating the colour-difference image.

Such convolutions can be used to aid, for example, the reduction of the effect on the end result, of the statistical noise appearing in images.

If the digital images are not in the same colour space in the initial situation, they can be converted to the same colour space prior to the calculation, preferably to a machine- independent colour space, such as the Lab space, by exploiting suitable ICC profiles. Individual Delta E values can be weighted using various parameters, in order to obtain different quality values. The weighting parameter can take into account, for example, people's subjective experience of the degree of disturbance of the Delta E values. The choice of the parameters can be based, for example, on a tone scale, an image subject, or the location of the subject within the image.

As stated above, the method is particularly suitable for situations, in which at least one of the images being compared is initially a tangible graphical product, for example, on paper. It must then be digitalized before the calculation of the reference variable or variables. Digitalization can be performed using, for example, a spectrum camera, a digital camera, or a scanner. The illumination environment prevailing during digitalization should be precisely defined and stabilized. Digitalization is preferably made into a spectrum form, so that the lighting values of the image can be digitally edited.

In a typical case, for example, in many claims cases and when printing according to a contract proof, both images are initially on paper. In that case, they are digitalized in the same conditions. The present method is particularly advantageous if the images have been produced using different methods, for example, one image has been produced using a proofing device or computer printer and the other image using a printing press. Microscopic differences in the products, arising from differences in the colour formation in different technologies, can be eliminated by using digital resolution selection, or by filtering digitally.

The term paper-based product refers not only to papers, but also to board and cardboard that can be used for printing or outputting. The printing and proofing methods used can be any processes whatever known in the art, including mechanical methods, such as relief printing, flat printing, offset, gravure printing, flexography, and serigraphy, as well as digital methods, such as electro-photography and laser and inkjet methods.

If an error is detected in the images on the basis of the comparison, an attempt can be made to correct it in several ways. In addition to altering the print profile, other printing parameters too, such as the type of printing ink, the binder, and the paper grade can affect the end result of a printed product. Altering any of these may correct the colour error between the proof and the printed product.

According to one preferred embodiment, the ability of the proofing machine and/or printing press to reproduce colour scales is evaluated. In that case, the images preferably include at least one continuous colour scale placed in some geometric shape, which covers a specific tone area, preferably all the tones from unsaturated tone ... saturated tone. The unsaturated tone is typically white and the saturated tone could be, for example, green, red, cyan, magenta, yellow, black, or some other tone formed with their aid.

If the digital images being compared are in different colour spaces, they can be converted to the same colour space prior to comparison. The spectral properties of one or other image, or of both images, such as colour temperature, can be altered, for example, in order to simulate the intended print quality in different lighting conditions. The colour-difference image will reveal in which locations of the images and in which tones a change in illumination will produce the greatest effect. In the case of advertisements for example, it is important for an advertisement to have an illustrative appearance in different lighting conditions, for example, in 'white' daylight, as well as under a yellowish or greenish street lamp.

The images can also be aligned with each other digitally, if the images were originally unaligned.

On the basis of the corresponding Delta E values of several image elements, or of quality values derived from them, various characteristics, i.e. quality parameters, can be calculated. The characteristic can be, for example, the mean of the Delta E values of a specific image area or tone area. If several characteristics are calculated, they can be given varying weighting values in decision making. Generally, it is not sensible to define precise criteria other than for compact (entirely covering) colour surfaces, as in other cases the nature of the image (e.g., the tones or the other colours or patterns close to the location being examined) affect the experience of colours. For examples, colour differences in skin tones are noticed with greater sensitivity than in saturated colour tones. Thus, in a typical application, the colour-difference limits to be used in the final decision-making are set case-specifically, generally in co-operation with the customer of the service (e.g., in printing work, the customer sets the Delta E limit for changes to skin tone, logos, etc.). However, in highly developed automation at least an initial decision on the similarity of the first and second image can be made automatically when a quality parameter or parameters meet a specific criterion.

The following examples depict different cases, in which various embodiments of the invention can be exploited, as well as examples of methods for these cases.

Example 1. Comparison of a proof and printed product, e.g., in a reclamation case

The printed product and the proof are digitalized into a spectral form, using a spectrum camera and standardized lighting conditions. The digitalized image files are opened and a colour-difference image is calculated on their basis. On the basis of the colour-difference image the customer approves the Delta E limits that should be used in his products.

In such a case, the present method is applied as follows:

— a proof is taken, which is digitalized to create a first digital image,

— a printed product, which is produced with specific printing parameters and corresponds to the proof, is taken and digitalized, in order to create a second digital image,

— a variable, preferably several such variables, proportional to the magnitude of the difference in the colour values of at least some image elements corresponding to each other in the image plane of the first and second digital images, is calculated, and - if the variable or variables meet a pre-defined criterion, the printing parameters are altered, in order to produce a new printed product.

The printing parameters, which affect the colour world of the end product, comprise, for example, the active adjustment parameters of the printing press, such as the amount (density) of the printing-ink feed in offset printing, the print profile used, the type and binder of the printing ink, and the paper grade. Altering any one of these can correct a colour error between the proof and the printed product. The method according to this example can be applied correspondingly to a situation, in which the products being compared are an image original and a proof, or for example an image original and a printed product. In some situations, instead of altering the printing parameters the proofing parameters can be altered.

Example 2.

The effect of lighting conditions

The product is digitalized as in example 1. Because the data is in a spectral form, the Lab values can be calculated simulated for different illuminations. Thus the customer will get print quality of what the printed product (e.g., a catalogue) will look like in the light of an incandescent lamp. On the basis of the colour-difference image, the Delta E values are obtained relative to the desired basic lighting.

In such a case, the present invention is applied as follows: - a first digital image is taken,

— the first digital image is doubled, in order to create a second digital image,

- the colour parameters of the second digital image are altered,

— a variables, proportional to the magnitude of the difference in the colour values of several image elements corresponding to each other in the image plane of the first and second digital images, is calculated, and

- if the variables meet a pre-defined criterion, a printed product is produced on the basis of the first digital image.

Example 3. Quality evaluation of a printing process

In printing, the press's own ICC profile is used, which is based on the density identities of specific fully covered areas and on the point growth of raster surfaces. Even though these printing technology variables are correctly set, it is not possible to be certain that the printing process operates optimally in all tones. By printing the critical image subjects, digitalizing the printed image according to Example 1, and comparing the digitalized images with the original digital data, information is obtained as to how well or poorly critical tones such as skin tones, sky, grass green, etc. will be reproduced. In such a case, the present method is applied as follows:

- a first digital image is taken, in which there is a printing-technologically critical image subject,

- on the basis of the first digital image, a paper-based product is produced by outputting or printing, using correspondingly specific proofing or printing parameters,

- the paper-based product is taken and digitalized, in order to create the second digital image,

- variables proportional to the difference in magnitude of the colour values of several image elements corresponding to each in the image plane of the first and second digital images, and

- if the variables meet a predefined criterion, the proofing or printing parameters are altered.

The following example describes the application of the present method to the quality evaluation of a print profile, in a situation, in which the critical image subject comprises a colour scale. It should be noted that the method can also be applied to entirely digitally edited image subjects, without these necessarily needing to be printed or output.

Example 4,

Reproduction of the colour scales of a print profile

It is difficult to evaluate the evenness of a colour scale created by a specific print profile utilizing the traditional printing quality measures. As a point of departure, it can be assumed that a colour scale in the RGB space will be faultless. When the scale is converted to the CMYK space, points of discontinuity and actual colour errors will often appear in the colour scale. However, it is important for colour scales to succeed without problems in image subjects of many different types, such as sky, or the depiction of metal at the boundaries of light and shadow. By comparing printed colour scales and/or digitally generated colour scales with the aid of a colour-difference image produced with the aid of the present method the errors in the evenness of the colour scales will be brought out best.

Figures 3a - 3c show an example of the reproduction of colour scales. In Figure 3a, the colour scale is from dark to white in the RGB space, while in Figure 3b it is in the CMYK space. Figure 3c shows the colour-difference image created on the basis of the images. In Figure 3 c, black represents a small colour difference and white a large difference. The points of discontinuity appear clearly. In such a situation the print profile does not produce optimal quality in the colour scales.

Claims

Claims:

1. Method for determining the quality of print quality, characterized in that

- a first digital image, which represents the intended print quality in the graphical product, is taken

- a second digital image, which represents the achieved or planned print quality of the graphical product, is taken,

- a variable proportional to the magnitude of the difference in the colour values of at least some image elements corresponding to each other in the image plane of the first and second digital images, is calculated, and on the basis of the said variable, the similarity of the colour worlds of the first and second images is estimated.

2. Method according to Claim 1, characterized in that the said variable is calculated for all the image elements of the image plane, in order to create a third digital image with the aid of the variables and several variables are used when estimating the similarity of the colour worlds of the first and second image.

3. Method according to Claim 1 or 2, characterized in that the number of degrees of freedom of the said variable is less that the number of degrees of freedom of the colour values of the image elements of the first and second digital images.

4. Method according to any of the above Claims, characterized in that a first paper-based product, which comprises graphical patterns, is taken, and the graphical patterns of the first product are digitalized, in order to create the first digital image.

5. Method according to any of the above Claims, characterized in that a second paper- based product, which comprises graphical patterns, is taken and the graphical patterns of the second product are digitalized, in order to create the second digital image.

6. Method according to Claim 5 depending on Claim 4, characterized in that the graphical patterns of the first and second paper-based products are produced by different methods, typically the patterns of the first product by outputting and the patterns of the second product by printing, or vice versa.

7. Method according to any of Claims 4 - 6, characterized in that the digitalization is performed in standard lighting conditions, into a spectral form.

8. Method according to any of the above Claims, characterized in that the spectral properties, such as the colour temperature, of the second digital image are altered, in order to simulate the intended print quality, in different lighting conditions.

9. Method according to any of the above Claims, characterized in that the first and second digital images are converted to the same colour space by using correspondingly a first and second colour-conversion profile, if the images are originally in different colour spaces.

10. Method according to any of the above Claims, characterized in that the first and second images are aligned with each other, if the images are originally unaligned.

11. Method according to any of the above Claims, characterized in that several variables proportional to the difference in the magnitude of the colour values of the image elements corresponding to each other in the image plane of the first and second images are calculated, on the basis of which at least one quality parameter is calculated.

12. Method according to Claim 11, characterized in that the several image elements corresponding to each other in the image plane of the first and second images are selected according to the different tone or subject areas of the first and second digital images.

13. Method according to Claim 11 or 12, characterized in that at least two quality parameters are calculated on the basis of image-element sets that differ at least partly from each other.

14. Method according to any of Claims 11 - 13, characterized in that, if the quality parameter meets the predefined criteria, the similarity of the colour worlds of the first and second image is positive.

15. Method according to any of the above Claims, characterized in that the first and second digital image comprise one or more colour scales, which cover the tone range unsaturated tone...saturated tone.

16. Method according to any of the above Claims, characterized in that the evaluation of the similarity of the colour worlds of the first and second image comprises some of the following: evaluation of the similarity of colour reproduction, evaluation of the similarity of tonality, evaluation of the geometrical deviation of the colour world, mean value of the difference in magnitude of the colour values of a specific image or tone area.

17. Method according to any of the above Claims, characterized in that

- a proof is taken, which is digitalized to create a first digital image,

- a printed product, which is produced with specific printing parameters, corresponding to the proof is taken and digitalized, in order to create a second digital image,

- a variable, preferably several such variables, proportional to the magnitude of the difference in the colour values of at least some image elements corresponding to each other in the image plane of the first and second digital images, is calculated, and — if the variable or variables meet a pre-defined criterion, the printing parameters are altered, in order to produce a new printed product.

18. Method according to any of Claims 1 - 16, characterized in that

- a first digital image is taken, — the first digital image is doubled, in order to create a second digital image,

- the colour parameters of the second digital image are altered,

- a variables, proportional to the magnitude of the difference in the colour values of several image elements corresponding to each other in the image plane of the first and second digital images, is calculated, and - if the variables meet a pre-defined criterion, a printed product is produced on the basis of the first digital image. °

19. Method according to any of Claims 1 - 16, characterized in that - a first digital image is taken, in which there is a printing-technologically critical image subject,

— if the variables meet a predefined criterion, the proofing or printing parameters are altered.

20. Method for improving print quality, characterized in that — a paper-based proof is taken, which has been produced with specific proofing parameters and the proof is digitalized, in order to create a first digital image,

— a paper-based or digital image original or paper-based printed product corresponding to the image subject of the proof, and which has been produced with specific print parameters, and if necessary the image original is digitalized, in order to create a second digital image,

— variables proportional to the difference in magnitude of the colour values of the image elements corresponding to each other in the image plane of the first and second digital images, and

— if the variables meet a predefined criterion, a new paper-based proof or printed product is produced by using new proofing or print parameters.

21. Use of a method according to any of Claims 1 - 20 for adjusting the print profile of a printing press.

22. Use of a method according to any of Claims 1 - 20 for selecting the paper grade for a printing process.

23. Use of a method according to any of Claims 1 - 20 for adjusting the amount of the feed of printing ink or the amount of the feed of moisturizing water, or for selecting the type of printing ink, or the binder of the printing ink.