RU2333613C1 - Method of artwork verification with view to authorship - Google Patents

Abstract

FIELD: information systems.

SUBSTANCE: invention refers to methods of cultural items identification, particularly for identification of artwork. The method lies in generation of encrypted painting fragments data according corresponding palette colours of the artwork. Then the unique features of the artwork, belonging only to the specific author and available in every painting of this author, are determined using genetic belonging of initial colours due to determination of extinction coefficient (Ec) of overtones from the proportion Rc=Ec(R+G+B)/3, where Rc - total colour consisting of red - R, green - G, blue - B colours. Determination of the unique features of the analysed painting object of the suggested author is performed the same way as during determination of unique features of the original painting object. When identifying the analysed object, comparison of revealed unique features of the analysed painting object and the original painting object is performed according graphical characteristic curve between Ec and Rc, calculated in the most dark, light and median colours taking into account usual hardware deviation. After that conclusion about authorship is made.

EFFECT: enhancement of functional abilities.

10 dwg, 1 tbl

Description

The invention relates to methods for identifying cultural objects, in particular for the identification of paintings in order to identify the author and protect them from fakes.

Known methods for marking various objects in a variety of ways using, for example, markers, holograms, using built-in electronic circuits, etc. The disadvantage of known marking methods and devices is that they do not allow recording identification information, but only give an answer, whether an authentic document or not .

There are also known systems for identifying objects, which are based on unique, inimitable, inherent only to this sample features.

From US 5521984 A, dated 05/28/1996, a method and system for determining the authenticity of objects, such as original works of art through the use of unique, inherent signs, is known. The method includes placing the studied image object in equipment that provides recording of a microscopic image, selecting a micro-section on the subject under study, recording an enlarged microscopic image of unique features on the selected micro-section of the subject, and storing data in a central repository. Subsequently, if it is necessary to identify an object, the data is retrieved from their storage and a comparison is made of the corresponding images of the micro-sections of the stored and studied image object.

In addition, from US 6017218 A, dated January 25, 2000, a method for identifying painting works is known, which consists in comparing brush strokes or brush marks of a studied work with strokes of one or more known painting works.

A known method of identifying religious objects (RU 2205451 C1, dated 05.27.2003), which consists in comparing the obtained image of at least one area of the identifiable object with the registered image of the area, and when choosing each surface area, the coordinates of the point are randomly set , relative to which the shape and size of this surface area are specified.

Known methods only simplify the visual inspection of works of art, but they do not allow an accurate analysis of a painting to determine its belonging to a particular master.

The technical result is the development of technology that allows you to analyze the work of painting by color in order to identify the author and distinguish the author’s painting from any kind of fakes, and record them in the author’s data bank.

To do this, in the method of identifying paintings for their authorship and authenticity, which consists in the fact that prior to identifying the object under study, the paintings are examined by the well-known originals of paintings by certain authors for the presence of unique features unique to a particular author, they form an author database identified unique features, identify the studied object of a painting, the presence of unique features inherent only definitely to the author of the painting, and comparing the identified unique features of the object under study with the unique features inherent in a particular author, stored in the author's databank, the study of known originals of paintings by certain authors for the presence of unique features is carried out by obtaining an image of the original painting using color digital photographing and outputting image data to a computer monitor using computer frames studies, extract data from fragments of paintings with different color spots that make up the palette of the work for their study, encrypt data from fragments of paintings in the RGB profile in one of the ADOBE programs by assigning each tone from the range of color changes (R + G + B) / 3 corresponding seven-bit code from position 128 (2 ^7), form a set of encrypted data fragment relevant to the painting color palette painting works by separating them on the basic RGB color tones, identify the presence of unique features Arts Nogo paintings that are unique to a specific author paintings and expected in any work of the author due to genetic supplies raw callers, by determining the coefficient of extinction k _g overtones from the relationship Rc = k _r (R + G + B) / 3 where Rc - total Kohler, consisting of red - R, green - G, and blue - B colors, the identification of unique features of the studied object of the painting by the alleged author is carried out in the same way as when identifying unique features of the original work When identifying the object under investigation, they compare the identified unique features of the object under study with the paintings and the original according to the graphical dependences of k _g on R, calculated from the darkest, lightest and middle colors, taking into account the natural hardware error, and conclude that the painting is identified on the subject their authorship.

Figure 1 - color distribution within the boundaries of the region D without changes,

In Fig 2 - a continuous change of color along the plane within the boundaries of region D.

Figure 3 is a set of color spots.

Figure 4 is a structural diagram of a device that implements the claimed method.

Figure 5-10 - graphs of changes in k _g from R.

The annex (not published) reflects the research protocols of paintings of culture and presents the conclusions.

A device for identifying a painting that implements the claimed method comprises a means 1 for obtaining unique features of the original paintings, connected to the author's data bank 2, means for obtaining a photographic digital image 3 of the painting, the output signal of which is via a digital recording unit in the RGB image profile 4, encrypt the data of the fragments of the painting in the RGB profile in one of the ADOBE programs by assigning to each tone from the color range (R + G + B) / 3 the corresponding sem bit code of the 128 position (2 ⁷⁾ is supplied to the extracting unit 5 digital overtones test object paintings, whose output is connected to the input fixing unit 6 digital overtones generator set 7 basic RGB digital tones and digital overtones relative red R = k _g (R + G + B) / 3, green G = k _g (R + G + B) / 3 and blue B = k _g (R + G + B) / 3 colors with a given gradation, connected to an 8 digital memory Kohler with their overtones, the output of which is connected to the reference input of the block selection 5 digital overtones of the investigated object produced painting Ia, output block fixation digital overtones 6 received light image is connected with the computation unit 9, the relationship k _g of dark to k _g of light callers whose output is connected to the block identification of the test 10 works, the reference input of which is connected to the output of copyright bank 2 data. The device can be implemented on a computer with the program "ADOBE", in the RGB-profile.

The proposed method for identifying the author of a pictorial work of culture is based on Goethe's teaching on color, according to which theoretically any color (color, tone, semitone, brightness) can be composed of three primary colors: red, yellow and blue. (Hereinafter, everywhere, “primary yellow” is replaced by “additional I-degree green”, since the invention used a digital method for describing color in an RGB profile, although the present invention can be applied with some changes using all other profiles: CMYK, LAB and etc.). Based on the isosceles of the Goethe triangle, in its center should be an ideally gray point, including on a black and white scale (1 + 255) / 2 = 128 (gradation) of tones. [To date, in the digital description, it is accepted that the range of color changes lies in the range 0-255. However, “0” conflicts with Goethe’s teachings. Therefore, the author accepted (111) as the lower minimum limit of the color change scale for R, G, and B, and “1” for the black and white scale, respectively]. Then its total spectral density in the RGB profile on the color scale (in the center of the triangle) will be respectively equal to (128 + 128 + 128) / 3 = 128 tones (gradations). Each tone is encrypted with its own code. It follows that, theoretically, the spectrum of any color in the language of the generally accepted digital (matrix) description of color in the RGB profile can be represented as the arithmetic mean of the sum of its constituent colors: (R + G + B) / 3. Further, if in the RGB profile, by analogy with musical literacy, the color is described as an overtone (fundamental tone), combined with the first and second tones, then it is easy to notice that the contribution of the last two to the resulting color tone is to quench the fundamental tone, since the color brightness will always be lower brightness overtone. Therefore, in the case of red, yellow, brown, etc. tones for which Rс acts as an overtone, the color can be described by the equation: Rc> (R + G + B) / 3, - (lvl.1). For green - Gc> (R + G + B) / 3 and, finally, for blue - Bc> (R + G + B) / 3. All these inequalities will turn into identities, if we take into account the damping factor of the overtone in them: k _g . Those. in the case of red, yellow and other tones, for which R acts as an overtone, the equation will take the form: Rc = k _g [(R + G + B) / 3]; for green - Gc = k _g [(R + G + B) / 3] and, finally, for blue - Bc = k _g [(R + G + B) / 3]. (According to studies conducted by color psychologists, red color has an exciting effect on the human psyche. The brain of the latter, in turn, seeks to extinguish aggression R by substituting soothing green or blue-green to it, so the author suggested considering “Rc” as the overtone color value regardless of color).

4. Any painting can be represented as a set of color spots that do not have a layer thickness and are located in the image area. Then:

a) if within a certain continuous field of the plane D (Fig. 1) the color is continuous and does not change, then its mathematical description will be the following well-known equation: S = ∫∫dσ- (lv. 2), where S, theoretically, is the spectral density palettes on the plane D, Q (mg) or L (units of brightness);

b) if the color within the plane D (Fig. 2) begins to continuously change, then by linking these changes to the coordinate grid, we will get a regular histogram inside the Gaus curve that describes quantitative color changes:

However, the use of this method as well as the method of describing color in three coordinates, in the plane of solving the issue of identifying the author of a pictorial work of culture, is possible except for computers because of the extreme complexity of the palettes of the artistic images themselves.

5. In order to find out the possibility of using level 1 for this purpose, the author of the present invention first investigated the dependence of k _g on the color change of the tinting color within the color scales he modeled using a computer, as shown in the appendix (not published). The results of the study show the following.

but. If the color changes within one grayscale scale, then the dependence of k _g on the value of Rc is directly (inversely) proportional, thereby showing which direction the movement is in the opposite direction: thus, the “-” sign in cases of yellow halftones means that they change towards green, and in turn red towards yellow, and vice versa.

In addition, falling or ascending dependencies simply tell us - at the beginning or at the end of their scale, a quarter-ton moves: “+” - to the beginning, and “-” - to the end (see Figs. 5-10). Thus, knowing the value of k _g , the researcher can, without seeing the color, compose an idea of his midtones. It should be noted that this position may be violated by the equality k _g for quarters of different colors.

However, this misunderstanding is completely removed if we take into account the fact that for a painting researcher, k _{g is} logically attached to (R + G + B) / 3 and, therefore, to the color of the color.

b. If the color varies linearly on a scale within two or more semitones, then the dependence of k _g on the argument becomes polynomial in nature, thus, in full compliance with the law of “Transition of quantity into quality”.

at. Further, if the color does not change linearly, then on the graph we get harmonic series (waves), the completeness of the members of which will depend on the completeness of the circles (scales) along which it moves.

Finally, based on the points a, b, and c, it can be concluded that k _g in the color equation R = k _g [(R + G + B) / 3] can be used in as a functional characteristic of color and its spectral components.

6. Probably, “color irritations affect the whole body-spiritual organism of a person: orange and red are perceived as exciting; green - soothing, blue gives the impression of depth, and yellow - the joy ", etc. “Through experiments, psychologists prove that each person has his own, intrinsic, color scale, that in color he can express his mood, character traits, perceptions and feelings.” “Observations also showed that each person has his own attitude to color”, i.e. "He has a whole scale of favorite colors." (I. Itten. The art of color). At the same time, “since a person’s penchant for one color or another can change under the influence of material or spiritual factors, the idea arises that a personal color scale can be considered to a certain extent as an expression of individuality,” argue G. Freeling and K. Auer, basing his judgments on the theory of I. Goethe about color and lectures by I. Itten, (Man - color - space. G. Freeling, K. Auer. Moscow, Stroyizdat, 1973).

So, the scale discussed above is subject to change and, therefore, according to the inventor, cannot be the basis for the method of identifying a painter whose brush belongs to a particular work. On the other hand, in the course of his own research, the author found that each individual has his own genetic scales of half-tone (quarter-tone) ranges, which are inherent only to him, which, in fact, explains the uniqueness of subjective colors. Simply put, in response to a signal received from outside, the human brain, from the genetic files of color scales at its disposal, substitutes its speculative color to the first (color, color). Moreover, the proximity of the latter to the first depends on the innate sense of color of the recipient and on how accurately his eyes allow him to distinguish between gradations of the quarter-tones of what he saw. So, due to deterioration of vision, which leads to a narrowing of the range of vision of the latter, the recipient begins to make mistakes when reproducing them, while, of course, the very set of genetic scales of the subject remains unchanged. In general, the presence of a speculative color vision in a person in the absence of an external signal makes us think that he (the person) always uses exclusively his own visual rows, and the color from the outside is only an excuse or only a command given through the eyes to the brain, executing which the latter searches for in its annals, the color is the closest in hue and provides its internal vision to the user. Thus, the inventor believes that the difference between painters should be sought not only in their "favorite color scales" (which is so changeable!), But, first of all, in their genetic and, therefore, not changing from beginning to end their life's vision of halftones and quarter tones, because, during the study, it was found that the color data of the works of one artist differ from the same data of the other artist precisely the half (quarter) tone color scales in which they are written, and this difference is through the equation R = k _g [(R + G + B) / 3] can be expressed by the damping factor of the overtone R. 1. The digital method for describing color and the author's color equation are based on the invention: in the RGB profile R = k _g [(R + G + B) / 3] - (ur .1), where R, G and B are the colors and their brightness that form the color; k _g is the damping coefficient of the overtone R (see table).

3. Slides from fifteen paintings by artists such as I. Aivazovsky, I. Levitan, B. Kustodiev, V. Polenov, K. Kastalsky-Borozdin, etc., were used as material for computer research. In total, 53 studies were conducted in four colors: red, yellow, green and blue. In the course of studying painting, the inventor confirmed the assumption that he originally made that the unique characteristic of any colorist is his unique genetic set of semi (quarter) tone scales, which he uses (used) to create his creations. In turn, since within the same color, each grayscale scale has its own range of changes in k _g , therefore, in relation to the subject, we can assume that at its beginning it bears the function of the genetic component of level 1. Then, when the color moves from the first quarter to the next, k _g will be equal to: k _gene + Δk _g , where just Δ and reflects the dependence of qualitative on quantitative changes. However, a logical question arises here: how likely is the repeatability of the same tones in different works of the same master? Studies have shown that, firstly, if each grayscale scale consists of eight quarter tones, then at best artists use it for one second, and the scales themselves in the same color - also no more than three, four. This means that the range of shades distinguished by the colorist eye is very small and, on the contrary, the probability of their repetition is high.

Thus, due to the limited genetic choice of the painter, they will all be expected. So, when comparing two, at first glance, paintings of V. Polenov absolutely different in color, “Grandma’s Garden” and “Moscow Courtyard”, it is found that, for example, the bright green colors of the “Moscow Courtyard” are scattered in the dark green bushes of “Grandma’s Garden” ”, And the juicy ocher of autumn leaves of the“ Grandmother’s Garden ”we will find in the core of the daisies of the“ Moscow Courtyard ”, etc. But it is this fact, the fact of artists' desire for innovation, that greatly complicates the task of identification. Therefore, the author of this study, knowing that painters, to create the integrity of a work, as a rule, balance dark and light spots, and that sensory equipment (photo and computer) works on the same principle, proposed the following simplification of the “Method”: the author suggested that if on a slide from an artist’s painting one finds the darkest color of the same color and the brightest (light) color of the same color, then the dimming coefficient of their arithmetic mean color will be expected for any other work of this m Stera, but because genetic source supplies callers and damping coefficient arithmetic mean caramel as value functionally dependent on them, will be a distinctive feature of one painter from another. It doesn’t matter at all whether this color scheme was used by the master himself, or it exists only in the matrix, or it is generally virtual, as a calculated quantity, it will still be expected for this author. In most cases, colors similar to the median ones were found on the slides of the paintings examined by the author).

As an evidence of the viability of the proposed “Method” in the appendix (not published), the author gives the results of studies of paintings of the above painters, the graphic processing of which allows us to conclude that the proposed approach to solving the problem of identifying the author of a pictorial work of culture is true. (Since, according to the author, the “databases” in themselves in the near future will represent both scientific and commercial interests, he considered it necessary to hide the names of the works he studied, with the exception of the work of I. Levitan's “Golden Autumn” (No. 02) For which he apologizes).

Figure 5-10 presents the arithmetic mean colors and the corresponding k _{g of} various works of various authors obtained by the author during the study of the works of the above-mentioned masters. Their comparison in yellow colors showed:

1. a) The extinction coefficients of 1.4263 and 1.4267 of I. Aivazovsky’s paintings “05” and “08”, respectively, are identical within the margin of error and indicate that “05” and “08” belong to the same master; b) in “05” the painter used a wider range of yellow shades, therefore her actual k _g cf., tint = 1.3293 (see 05-1), which does not cancel the conclusion of “a”; c) this k _g is close in value to k _g = 1.3125 by I. Levitan (see 02–2) and k _g = 1.2901 by V. Polenov (see “14”). However, taking into account their original (author's) colors, it becomes obvious that they come from completely different grayscale scales and, therefore, belong to different authors.

2. All k _g , both the median and the original tints of the works “01” and “02” by I. Levitan are almost identical and, thus, confirm their belonging to one author. On the contrary, they differ from the _{kg of} I. Aivazovsky and V. Polenov, thereby indicating that “01” and “02” do not belong to the latter. An exception is the color 02-2 already mentioned in Clause 1, “C”, but it is not median and, therefore, may not be taken into account at all.

3. The works of V. Polenov according to their color data are very different from the similar characteristics of other authors studied in this work, and this difference is that the artist added much more whiteness to the colors in the “lights” than the latter, manifests itself in quantitative content of blue in them. Therefore, its middle color, similar in color and k _g value (see “14”) to colors of 05-2 and 02-2 of I. Ayvazovsky and I. Levitan, in fact, has nothing to do with them. On the other hand, V. Polenov, like no one else, loved innovation and in his new works sought not to repeat the colors he had already used. Their quarter-tone difference is what we observe when comparing his works. Nevertheless, it turned out that its middle colors and their damping coefficients are almost identical (see “13” and “15”) and confirm the author’s assumption made earlier regarding their expectation.

As for “14”, k _{g of} which is slightly different from the blanking coefficients “13” and “15”, this is explained by the too large difference in hardware errors in blue between the slides from these works (The research protocols of the corresponding paintings are given in the Appendix - are not published).

4. Unfortunately, the magnitude of the errors in photographing the works “10” and “11” by B. Kustodiev did not allow us to adjust their databases with the databases of other authors. However, the relative hardware error between them turned out to be comparable and allowed identification of the ownership of these paintings by one author. The study led to a negative result, which was to be expected in this case. (Slides were made with printed matter.)

Comparison of k _{g of the} middle tones for green tones in general confirms the conclusions made above for yellow tones.

Thus, as shown above, k _{g of the} middle color is a distinctive characteristic of the painter. Nevertheless, k _g per se, being very good guidelines for the researcher, does not allow to take into account the natural errors and, therefore, does not provide the presentability of the “Method”, therefore, comparison of authors only on them can lead the expert to serious errors, since the latter (natural errors) just manifest themselves not in the middle, but along the borders of the scales.

Based on the foregoing, the author of the invention considers it necessary to use a graphical comparison of the results for three k _g : the darkest, lightest and intermediate (middle) colors. From the graphs given in the Appendix (not published) errors are very clearly visible to which the expert will inevitably come, due to their neglect of natural errors (for example, the paintings “05” and “08” do not belong to the same master!).

So, according to the results of the graphic examination of the work “05” and “08” belong to the brush of the same master; distinct works "01" and "02" belong to the brush of the same, but different master. The works “13”, “14” and “15” belong to the brush of the same, but the third master. B. Kustodiev’s belonging to the works “10” and “11” has not been established due to the high hardware error of photography provided to the researcher. And, finally, it was established that all four sketches of K. Kastalsky-Borozdin: “Dramatic Sunset”, “Silver Day”, “After the Rain” and “Morning on the Psel River” were written by one master. Finally, the “Method” described above proved its applicability with the aim of identifying the author of a pictorial work of culture and also showed that it has high resolution comparable to the sensitivity of photo-and-sensory equipment.

In conclusion to all of the above, I would like to emphasize that one of the key issues in solving the problem is to take into account the hardware and natural errors associated with the deterioration of artists' vision, as well as the inconstancy of lighting the artist’s workplace, which, in turn, leads to the inconstancy of the ranges of visible quarter-tones along their borders, as well as the representativeness of the selection of colors. The author has developed appropriate methods for accounting for errors, however, since they themselves have the sign of “novelty,” he (the author) considered it possible not to disclose them in the description of the present invention. Representativeness was ensured by the choice of colors with the closest possible values of the actual overtones, as it is easy to notice in the materials shown in figure 5-10.

Claims (1)

A method of identifying paintings for their authorship and authenticity, which consists in the fact that prior to identifying the object under study, the paintings are examined by the well-known originals of paintings by certain authors for the presence of unique features that are unique to a particular author, and an author database of identified unique characters is formed , identify the studied object of a work of painting the presence of unique features inherent only to a particular author pr painting works and compare the revealed unique features of the object under study with the unique features inherent in a particular author, stored in the author's data bank, characterized in that the research of known originals of paintings by certain authors for the presence of unique features is carried out by obtaining an original image of the painting using color digital photographing and outputting image data to a computer monitor screen, through a computer the adrenals highlight the data of the fragments of the painting with different color spots that make up the palette of the work for their study, encrypt the data of the fragments of the painting in the RGB profile in one of the ADOBE programs by assigning each tone from the color range (R + G + B) / 3 corresponding from 128 gradations of tones, form a set of encrypted data of fragments of the painting according to the corresponding colors of the palette of the work by dividing them into the main color tones of RGB, identify the presence of unique signs of the work of art inherent only to a certain author of a painting and expected in any work of this author due to the genetic affiliation of the original colors, by determining the coefficient of damping overtones from the ratio Rc = k _g (R + G + B) / 3, where Rc is the total color, consisting of red - R, green - G and blue - In colors, the identification of unique features of the studied object of the painting by the alleged author is carried out in the same way as when identifying the unique features of the original painting, when identifying katsii test object is compared identified by unique characteristics of the test object paintings and original on the graphic dependences K _r from Rc, calculated in the dark, the light and middle Kohler with the natural hardware error, and conclude identification paintings for their authorship.

Images