Method for determining at least one colorimetric parameter characterizing a cosmetic composition, associated electronic determination device, and computer program product

Abstract

The present invention relates to a method for determining at least one colorimetric parameter characterizing hair (12), in particular a cosmetic composition (11) for dyeing hair, comprising a learning phase including the steps of receiving a set of feasibility constraints, collecting a set of training data, and training an artificial intelligence model using the set of training data to obtain a trained model. The method further includes a processing phase including the steps of obtaining a data item under test and applying the trained model to the data item under test to determine at least one associated colorimetric parameter, wherein the data item under test complies with a set of feasibility constraints.

Description

[Technical Field]

[0001] A first invention relates to a method for determining at least one colorimetric parameter. The invention also relates to an associated computer program product and an electronic determination device.

[0002] The first invention relates to the field of cosmetic products, preferably cosmetic compositions for dyeing hair, in particular hair. [Background technology]

[0003] More generally, a "cosmetic product" is a product as defined in Regulation EC 1223 / 2009 of the European Parliament and of the Council of 30 November 2009 on cosmetic products.

[0004] The goal of the cosmetics industry is to enhance the experience of its consumers. In particular, there is a growing trend to offer products that are increasingly adapted to the specific needs and characteristics of users. This trend is commonly referred to as "customization."

[0005] Customization of cosmetic products and services can concern any part of the human body, but is particularly important for exposed body parts such as the face (make-up or care products) and, where applicable, the hair or beard (e.g., coloring products).

[0006] In the field of cosmetic compositions for dyeing hair, and in particular hair, new cosmetic compositions are regularly developed in order to adapt to user demands.

[0007] These new cosmetic compositions are intended, in the first case, to achieve a new color or a new visual effect on the hair to which they are applied. In the second case, they are intended to achieve a previously known color, but comprise ingredients not previously used for this purpose. This second case is particularly important when some ingredients are difficult to source, involve high costs, or pose environmental risks.

[0008] However, obtaining new cosmetic compositions results from a very long and complicated process.

[0009] First, expert chemists develop cosmetic composition prototypes that are derived from the experience of expert chemists and are governed by several chemical principles.

[0010] These prototypes are then individually tested in the laboratory to ensure that the visual effects obtained meet the expectations of expert chemists.

[0011] Occasionally, and unpredictably, the visual effects associated with newly developed prototypes do not meet expectations, and therefore it is not possible to take advantage of the work of the expert chemists who developed these prototypes.

[0012] Therefore, there is a need for tools that make it possible to simplify the development of new cosmetic compositions. Summary of the Invention [Means for solving the problem]

[0013] The first invention proposes to overcome this problem by using artificial intelligence to aid in the development of new cosmetic compositions.

[0014] The first invention relates to a method for determining at least one colorimetric parameter characterizing a cosmetic composition for dyeing hair, in particular hair, the method being carried out by an electronic determination device and comprising a training phase, the training phase comprising: - receiving a set of feasibility constraints for a cosmetic composition; - collecting a set of training data, each training data item being specific to a cosmetic composition that complies with the set of received feasibility constraints; Each training data item is A set of sizes representing the amount of each ingredient in the cosmetic composition; the value of at least one colorimetric parameter associated with the cosmetic composition; and - training the artificial intelligence model based on the set of training data to obtain a trained model; Including, The method further includes a processing phase, the processing phase comprising: - obtaining test data items relating to the cosmetic composition under test, wherein the cosmetic composition under test complies with the set of feasibility constraints received during the training phase; the test data items comprise a set of magnitudes representing the amount of each ingredient in the cosmetic composition being tested; - applying the trained model to the acquired test data items to determine a value of at least one colorimetric parameter of the cosmetic composition under test; The present invention relates to a method, comprising:

[0015] Thanks to the artificial intelligence model, it is possible to quickly and easily determine at least one colorimetric parameter associated with each cosmetic composition under test.

[0016] Furthermore, the fact that each cosmetic composition complies with a set of constraints makes it possible to benefit from the experience of expert chemists and to avoid unnecessarily testing cosmetic compositions that will not be viable.

[0017] According to a particular embodiment, the determination method according to the first invention comprises one or several of the following characteristics, taken alone or in any technically possible combination: - each component of the test data item for the cosmetic composition under test is selected from a predefined list of components; - during the obtaining step of the processing phase, several test data items are obtained, each test data item being specific to a respective cosmetic composition under test; During an application step of the processing phase, the trained model is applied to each acquired test data item to determine a value of at least one colorimetric parameter associated with said test data item; The step to be acquired in the processing phase is receiving a number N corresponding to the number of cosmetic compositions under test, the number N being greater than or equal to 2; generating N test data items, each test data item relates to a cosmetic composition whose ingredients are selected from a predefined list of ingredients, such that the cosmetic composition associated with each test data item is within a reduced feasible cosmetic composition space; the reduced feasible cosmetic composition space comprises only cosmetic compositions that comply with a set of feasibility constraints; N test data items represent the reduced feasible cosmetic composition space; and - during the acquiring step of the processing phase, the step of generating N test data items of the acquiring step includes a step of calculating, for each test data item, the magnitude of the N test data items such that the distance between said test data item and other test data items is maximal; each magnitude of each test data item represents the amount of the ingredient in the corresponding cosmetic composition being tested; - the predefined list of ingredients comprises one or more ingredients of a first type and one or more ingredients of a second type; the or each component of the first type is a base; the or each component of the second type is a coupler; the set of cosmetic composition feasibility constraints includes the following constraints: the ratio between the amount of the first type of component and the amount of the second type of component is between a first threshold value and a second threshold value, The total amount of the cosmetic composition is less than the third threshold value; The amount of each ingredient in the cosmetic composition is less than the fourth threshold value; The number of ingredients in the cosmetic composition is less than the fifth threshold and - at least one colorimetric parameter is a triplet of values ​​characterizing the color of the hair after applying the cosmetic composition, or the color fading value of the cosmetic composition after washing the hair, or A selectivity value that characterizes the color difference between the root and tip of the hair. and - during a collection step of the training phase, the set of training data is collected, at least in part, via a sensor capable of measuring at least one colorimetric parameter; - the processing phase further includes sending the or each colorimetric parameter and associated test data item to a display screen for displaying a rendering for the or each cosmetic composition under test on the display screen, the or each rendering representing application of the cosmetic composition under test to hair; - the treatment phase further comprises the step of preparing at least one sample of the cosmetic composition under test for application on hair to verify the value of at least one colorimetric parameter.

[0018] The first invention also relates to a computer program product having stored thereon a computer program comprising program instructions, the computer program being loaded onto a data processing unit and performing such a method when the computer program is implemented on the data processing unit.

[0019] The first invention also relates to an electronic device for determining at least one colorimetric parameter characterizing a cosmetic composition for dyeing hair, in particular hair, comprising: It also relates to an electronic device, wherein the electronic decision-making device is capable of implementing such a decision-making method.

[0020] The first invention also relates to a readable information medium on which a computer program product with program instructions is stored, the computer program being loaded onto a data processing unit and which performs such a determination method when the computer program is implemented on the data processing unit.

[0021] Other characteristics and advantages of the first invention will become apparent on reading the following description of an embodiment of the invention, given purely by way of example and with reference to the drawings, in which: [Brief explanation of the drawings]

[0022] [Figure 1] FIG. 1 is a schematic diagram of the concept of the first invention. [Figure 2] 1 is a schematic diagram of an electronic decision-making device according to a first invention; [Figure 3] 3 is a flowchart of a decision-making method according to a first invention, implemented by the electronic decision-making device of FIG. 2; [Figure 4]FIG. 1 is a two-dimensional schematic representation of a reduced viable cosmetic composition space. [Figure 5] FIG. 1 is a schematic diagram of the second inventive concept. [Figure 6] FIG. 2 is a schematic diagram of an electronic decision-making device according to a second invention. [Figure 7] 7 is a flowchart of a decision-making method according to a second invention, implemented by the electronic decision-making device of FIG. 6; [Figure 8] FIG. 1 is a two-dimensional schematic representation of a reduced viable cosmetic composition space. [Figure 9] FIG. 8 is a two-dimensional schematic representation of the determination method shown in FIG. 7. [Figure 10] FIG. 10 is a schematic diagram of the third inventive concept. [Figure 11] FIG. 10 is a schematic diagram of an electronic adaptive device according to a third invention. [Figure 12] 12 is a flowchart of an adaptation method according to a third invention, implemented by the electronic adaptation device of FIG. [Figure 13] 13 is a flowchart of the steps of the adaptation method of FIG. 12. [Figure 14] FIG. 1 is a two-dimensional schematic representation of a reduced viable cosmetic composition space. [Figure 15] 14 is a two-dimensional schematic representation of the sub-steps shown in FIG. 13 of the adaptation method shown in FIG. 12. [Figure 16] 13 is a diagrammatic representation of the steps of the adaptation method shown in FIG. 12. DETAILED DESCRIPTION OF THE INVENTION

[0023] In FIG. 1, an electronic device 10 for determining at least one colorimetric parameter associated with a cosmetic composition 11 is depicted.

[0024] 1, the cosmetic composition 11 is a cosmetic composition for dyeing hair 12. Preferably, the cosmetic composition 11 is capable of being applied to hair 12 of a user 13 to dye said hair 12.

[0025] The cosmetic composition 11 comprises multiple ingredients 14 that are capable of interacting with each other and with the hair 12 of a user 13 to color it.

[0026] Preferably, during this interaction, component 14 of cosmetic composition 11 begins by bleaching hair 12. Then, component 14 of cosmetic composition interacts with the bleached hair to fix a pigment of the selected color, thus forming dyed hair 12*.

[0027] Ingredient 14 is present in cosmetic composition 11 in the form of, for example, a powder, a gel, an emulsion, or an oil.

[0028] The electronic determination device 10 is configured to determine the value of at least one colorimetric parameter characterizing the color of the hair 12 after application of the cosmetic composition 11 using a set of magnitudes representing the amount of each ingredient 14 in the cosmetic composition 11.

[0029] Referring to FIG. 2, the decision device 10 comprises a processing unit 15 .

[0030] The decision device 10 optionally further comprises a display screen 16 and / or a unit 17 for producing the cosmetic composition 11 .

[0031] The processing unit 15 comprises, for example, a calculator that interacts with a computer program product, for example, the processing unit 15 is a computer.

[0032] The computer comprises, for example, a processor comprising a data processing unit, a memory, an information medium reader, and optionally a human-machine interface.

[0033] The computer program product includes an information medium.

[0034] The information medium is typically a computer-readable medium by a data processing unit. The readable data medium is a medium adapted to store electronic instructions and capable of being coupled to a computer system bus.

[0035] For example, the information medium is a USB flash disk, a floppy disk or flexible disk ("floppy disk"), an optical disk, a CD-ROM, a magneto-optical disk, a ROM memory, a RAM memory, an EPROM memory, an EEPROM memory, a magnetic card, or an optical card.

[0036] A computer program comprising program instructions is stored on an information medium.

[0037] The computer program can be loaded onto the processing unit 15 and, when implemented on the processing unit of the computer, is adapted to perform a method for determining at least one colorimetric parameter, such a determination method being described herein below.

[0038] The operation of the electronic device 10 implementing a method for determining at least one colorimetric parameter will now be described with reference to the flow chart of FIG. 3 as well as the example of FIG.

[0039] The determination method includes a training phase 100 .

[0040] The training phase 100 includes receiving 110 a set of feasibility constraints for the cosmetic composition 11 .

[0041] Preferably, each cosmetic composition 11 comprises at least one ingredient 14 of a first type and at least one ingredient 14 of a second type.

[0042] Each component 14 of the first type is, for example, a base, and each component 14 of the second type is, for example, a coupler.

[0043] Preferably, each component 14 is selected from a predefined list of components.

[0044] Preferably, the set of feasibility constraints for the cosmetic composition 11 includes the following constraints: - the ratio between the amount of the first type of component 14 and the amount of the second type of component 14 is between a first threshold value and a second threshold value; - the total amount of the cosmetic composition 11 is less than the third threshold value; - the amount of each ingredient 14 in the cosmetic composition 11 is less than a fourth threshold value; - the number of ingredients 14 in the cosmetic composition 11 is less than the fifth threshold The present invention is provided with one or more of the following:

[0045] Advantageously, the set of constraints comprises each of the above constraints.

[0046] These constraints are optionally derived from chemical principles that ensure the viability of the cosmetic composition 11. The expression "viability of the cosmetic composition 11" is understood herein to mean that a composition that meets the set of constraints is effective and / or does not pose a risk to the user 13 and / or complies with product specifications and is preferably environmentally friendly.

[0047] The training phase 100 further includes a step 120 of collecting a set of training data 18 .

[0048] Each training data item 18 in the set of training data 18 is unique to a cosmetic composition 11 that complies with the set of feasibility constraints received during the receiving step 110 .

[0049] Each training data item 18 comprises a set of magnitudes representing the amount of each ingredient 14 in the cosmetic composition 11 and a value of at least one colorimetric parameter associated with the cosmetic composition 11 .

[0050] Each magnitude is, for example, the amount of the corresponding ingredient 14 in the associated cosmetic composition 11 .

[0051] Each amount of component 14 is optionally the amount of material in moles, the mass of the component, the volume of the component, the mass percentage of component 14 in cosmetic composition 11, or the volume percentage of component 14 in cosmetic composition 11.

[0052] The or each colorimetric parameter characterizes a cosmetic composition 11 for coloring hair 12 .

[0053] Preferably, each colorimetric parameter represents a visual effect on the hair 12 to which the cosmetic composition 11 is applied.

[0054] Advantageously, at least one colorimetric parameter is a triplet of values ​​characterizing the color of the hair 12 after applying the cosmetic composition 11, and / or - the color fade value of the cosmetic composition 11 after washing the hair 12, and / or - a selectivity value characterizing the color difference between the root and tip of the hair 12 Equipped with.

[0055] Preferably, each training data item 18 comprises each of the above values.

[0056] The value triplet is, for example, a CIE L*a*b* color space value triplet.

[0057] The CIE L*a*b* color space, often abbreviated as CIELAB, is a color space for surface colors defined by the International Commission on Illumination (CIE) in 1976. It is based on the CIE XYZ system of evaluation, abandoning linearity to more accurately highlight differences between colors perceived by the human eye. In this model, three dimensions characterize a color: lightness (L*), derived from luminance (Y) in the XYZ evaluation; and two parameters, a* and b*, which describe the color difference from the color of a gray surface with the same lightness, as chrominance. The definition of gray, uncolored, and achromatic surfaces implies that the composition of the light illuminating a colored surface is explicitly stated. This light source is often a daylight source corresponding to the D65 normalization standard.

[0058] Also in this case, the triplet of values ​​comprises a lightness value L* and two color difference values ​​a*, b* from a gray surface color having the same lightness.

[0059] Any other color space may be envisaged.

[0060] Alternatively, the triplet of values ​​is an RGB triplet, which then comprises a value R for red, a value G for green and a value B for blue.

[0061] More generally, any representation that makes it possible to relate spectra can also be used in this context.

[0062] Hair color is not equal at all points on the hair, in particular the ends of the hair 12 are generally lighter than the roots, therefore the selectivity value is an important parameter.

[0063] Preferably, during the collection step 120 of the training phase 100, the set of training data 18 is collected, at least in part, via a sensor capable of measuring at least one colorimetric parameter.

[0064] The sensor is, for example, a spectrocolorimeter.

[0065] According to one embodiment, the sensor is capable of capturing an image of a zone of hair of an individual and extracting colorimetric measurements from the captured image.

[0066] Preferably, each training data item 18 is derived from a cosmetic composition 11 that is feasible, ie, conforms to a set of feasibility constraints.

[0067] Preferably, each training data item corresponds to a cosmetic composition that is commercially available, and therefore, prior to marketing of said cosmetic composition, the colorimetric parameters of which have been measured during testing on hair 12.

[0068] Further, optionally, the cosmetic compositions associated with the training data are pre-selected by an expert chemist from all commercially available cosmetic compositions.

[0069] The training phase 100 further includes a step 130 of training the artificial intelligence model using the set of training data 18 to obtain a trained model.

[0070] Preferably, the artificial intelligence model comprises at least one of the following models: a support vector machine, a random forest, a gradient boosting mechanism, or Kriging, also known as Gaussian process regressor.

[0071] In each of the above cases, the artificial intelligence model comprises adjustable parameters, and during training step 130, each adjustable parameter is adjusted such that, when a respective set of magnitudes of training data items 18 is given as input, the trained model provides as output one or more colorimetric parameters that are substantially equal to the colorimetric parameters of the training data items 18.

[0072] The determination method further includes a processing phase 200 during which the trained model is used to determine colorimetric parameters associated with one or more cosmetic compositions under test.

[0073] The processing phase 200 includes a step 210 of obtaining test data items 19 relating to the cosmetic composition 11 under test.

[0074] For each cosmetic composition 11 under test, the test data items 19 comprise a set of magnitudes representing the amount of each ingredient 14 in the cosmetic composition under test.

[0075] Preferably, the ingredients 14 of each cosmetic composition 11 under test are selected from the predefined list of ingredients 14 described above.

[0076] Each cosmetic composition 11 under test conforms to a set of feasibility constraints received during the training phase 100 .

[0077] The obtaining step 210 advantageously includes receiving a number N corresponding to the number of cosmetic compositions 11 under test. Optionally, the number N is equal to or greater than 2. Preferably, the number N is greater than 100, for example, equal to 150.

[0078] The obtaining step 210 further advantageously includes generating N test data items 19, each test data item 19 relating to a cosmetic composition 11 whose ingredients 14 are selected from a predefined list of ingredients 14, such that the cosmetic composition 11 associated with each test data item 19 is within a reduced feasible cosmetic composition space 20. The reduced feasible cosmetic composition space 20 comprises only cosmetic compositions 11 that comply with a set of feasibility constraints. The test data items 19 represent the reduced feasible cosmetic composition space 20.

[0079] The expression "test data items 19 representing the reduced space 20" is understood in this specification to mean that the test data items 19 are appropriately selected in the reduced space 20 so as to substantially cover the entire reduced space 20.

[0080] 4 shows, via a two-dimensional block diagram, a set 22 representing all possible cosmetic compositions 11, and a reduced space 20. It is clear that this set 22 and this space 20 cannot in fact be represented two-dimensionally, but rather can be represented in a number of dimensions equal to the number of ingredients 14 in the predefined list of ingredients.

[0081] In Figure 4, a set 22 of all possible cosmetic compositions 11 corresponds to the map in Figure 4. As can be seen in Figure 4, a reduced space 20 is included in said set 22 because reduced space 20 comprises only those cosmetic compositions 11 from set 22 that meet the feasibility constraints.

[0082] 4, the test data items 19 can be seen in a reduced space 20. For example, ten test data items 19 are represented in FIG.

[0083] Preferably, generating the N test data items 19 includes calculating, for each test data item 19, the size of the N test data items 19 such that the distance between said test data item 19 and the other test data items 19 is maximum.

[0084] Each magnitude of each test data item 19 represents the amount of ingredient 14 in the corresponding cosmetic composition 11 under test.

[0085] Generating N test data items 19 includes calculating a magnitude among the N sets of magnitudes where the distance between each pair of the sets of magnitudes is greatest. Each magnitude in each set of magnitudes represents an amount of an ingredient 14 in the cosmetic composition 11 associated with that set of magnitudes, such that the cosmetic composition 11 complies with a set of constraints. For example, test data item 19 has a first magnitude equal to 4 moles of a first ingredient 14A and 6 moles of a second ingredient 14B.

[0086] Advantageously, each test data item 19 is a vector comprising a coefficient for each component 14 of a predefined list of components.

[0087] Each coefficient represents the amount of said ingredient 14 in the respective cosmetic composition 11 of the associated vector.

[0088] The distance between two test data items 19, i.e., between two vectors, is defined by an algebraic norm, for example, according to the following formula:

[0089]

number

[0090] where X and Y are two sets of size P, x i,i=1,...,P and y i,i=1,…,P are the respective sizes of the sets X and Y, √ is the square root function, Σ is the summation operator.

[0091] The N vectors are calculated, for example, by applying a design of experiments technique, also known as the DoE technique. Advantageously, the DoE technique implements an SFD (space filling design) algorithm, which makes it possible to ensure that the selected vectors represent the reduced space 20, even when using a reduced number of vectors. Indeed, when the number of vectors is low, the assumption of the law of large numbers is not sufficiently fulfilled to allow a random distribution of vectors to represent the reduced space 20.

[0092] In FIG. 4, it can be seen that the test data items 19 are spaced apart from one another to maximize the distance between the test data items 19.

[0093] Alternatively, the obtaining step 210 only involves collecting one or more test data items 19 selected by an operator of the decision device 10. In this case, the operator selects one or more cosmetic compositions by selecting, for each cosmetic composition, ingredients 14 from a predefined list of ingredients and the amount of each of said ingredients 14. When selecting, the operator ensures that each selected cosmetic composition 11 complies with a set of constraints.

[0094] The processing phase 200 further includes a step 220 of applying the trained model to the acquired test data items 19 to determine at least one colorimetric parameter of the or each associated cosmetic composition 11.

[0095] Preferably, each test data item 19 is sequentially fed to the trained model during the applying step 220. The trained model then determines, for each test data item 19, the associated colorimetric parameters.

[0096] Advantageously, during the applying step 220 , the cosmetic composition 11 and colorimetric parameters associated with each test data item are stored in one or more memories of the processing unit 15 .

[0097] Additionally, optionally, processing phase 200 further includes step 230 of sending the or each colorimetric parameter and associated test data item to display screen 16. Display screen 16 then displays a rendering representing the application of the cosmetic composition to hair 12 for each test data item.

[0098] Preferably, during the sending step 230, the display screen displays an image of the hair 12 sample to which the cosmetic composition associated with each test data item has been applied, each image being calculated using the colorimetric parameters.

[0099] If the colorimetric parameters comprise a triplet of values ​​characterizing the color of hair 12 after application of cosmetic composition 11, the display screen displays a first image of a sample of hair 12 to which the cosmetic composition associated with each test data item has been applied.

[0100] If the colorimetric parameters further comprise a color fading value of the cosmetic composition 11 after washing the hair 12, the display screen 16 further displays, for example, a second image of the same sample of hair 12 after washing the hair 12, calculated using the color fading value.

[0101] If the colorimetric parameters additionally or instead comprise a selectivity value, the display screen displays a zoomed third image of the sample of hair 12, for example showing the color difference between the root and tip of hair 12.

[0102] According to another optional addition, the processing phase 200 further comprises a step 240 of producing at least one sample of the cosmetic composition 11 under test. The at least one sample is preferably produced by a unit 17 for producing the cosmetic composition 11.

[0103] For example, during manufacturing step 240, samples of cosmetic composition 11 associated with each test data item are manufactured.

[0104] Alternatively, the manufacturing step 240 includes receiving from the operator a selection of one or more cosmetic compositions 11 to be manufactured from the cosmetic compositions 11 under test 11. The manufacturing unit 17 then manufactures only samples of the cosmetic compositions 11 selected by the operator.

[0105] A sample of the manufactured cosmetic composition 11 is intended to be applied to hair 12, preferably a tress of hair 12, in order to verify the colorimetric parameters.

[0106] Thus, following application of each sample on hair 12 , colorimetric parameters are measured using the same sensors used to form training data 18 .

[0107] Alternatively, the processing phase 200 does not include the sending step 230 and / or the manufacturing step 240 .

[0108] Alternatively, each cosmetic composition 11 can be applied to any type of hair 12, not just hair, for example, to the eyelashes, eyebrows, or beard of the user 13.

[0109] Using the determination method described above, the development of new cosmetic compositions 11 is accelerated because it is possible to digitally determine the colorimetric parameters associated with the cosmetic composition 11 without having to physically test the cosmetic composition 11.

[0110] Furthermore, this method allows a large number of cosmetic compositions to be tested quickly, making it possible to obtain a color that is more adapted to the user's desires.

[0111] Furthermore, the fact that the test data items 19 represent a reduced space 20 makes it possible to discover new, previously unsuspected cosmetic compositions 11. Furthermore, this makes it possible to obtain a substantial idea of ​​the different possibilities for cosmetic compositions 11 that fit a set of constraints.

[0112] All the variations and optional additions described above can be combined with each other.

[0113] Next, a second aspect of the invention will be described with reference to FIGS.

[0114] The second invention relates to a method for determining a target cosmetic composition for dyeing hair, in particular hair, according to at least one target colorimetric parameter. The second invention also relates to an associated computer program product and an electronic determination device.

[0115] The second invention relates to the field of cosmetic products, preferably cosmetic compositions for dyeing hair, in particular hair.

[0116] More generally, a "cosmetic product" is a product as defined in Regulation EC 1223 / 2009 of the European Parliament and of the Council of 30 November 2009 on cosmetic products.

[0117] The goal of the cosmetics industry is to enhance the experience of its consumers. In particular, there is a growing trend to offer products that are increasingly adapted to the specific needs and characteristics of users. This trend is commonly referred to as "customization."

[0118] Customization of cosmetic products and services can concern any part of the human body, but is particularly important for exposed body parts such as the face (make-up or care products) and, where applicable, the hair or beard (e.g., coloring products).

[0119] In the field of cosmetic compositions for dyeing hair, and in particular hair, new cosmetic compositions are regularly developed in order to adapt to user demands.

[0120] These new cosmetic compositions are intended, in the first case, to achieve a new color or a new visual effect on the hair to which they are applied. In the second case, they are intended to achieve a previously known color, but comprise ingredients not previously used for this purpose. This second case is particularly important when some ingredients are difficult to source, involve high costs, or pose environmental risks.

[0121] However, obtaining new cosmetic compositions results from a very long and complicated process.

[0122] First, expert chemists develop cosmetic composition prototypes that are derived from the experience of expert chemists and are governed by several chemical principles.

[0123] These prototypes are then individually tested in the laboratory to ensure that the visual effects obtained meet the expectations of expert chemists.

[0124] Occasionally, and unpredictably, the visual effects associated with newly developed prototypes do not meet expectations, and therefore it is not possible to take advantage of the work of the expert chemists who developed these prototypes.

[0125] Furthermore, it may be desirable to have several cosmetic compositions that produce the same color or visual effect, thereby preventing future problems due to the unavailability of some ingredients while ensuring that demand from users seeking to obtain said color or visual effect remains satiable.

[0126] Therefore, there is a need for tools that make it possible to simplify the development of new cosmetic compositions.

[0127] To this end, the second invention relates to a method for determining a target cosmetic composition for dyeing hair, in particular hair, according to at least one target colorimetric parameter, comprising: The method is performed by an electronic decision-making device and includes a processing phase, the processing phase comprising: - collecting at least one target colorimetric parameter; - obtaining a set of cosmetic compositions under test, each cosmetic composition under test comprising an ingredient; - obtaining at least one colorimetric parameter associated with each cosmetic composition under test; - filtering the cosmetic compositions under test according to at least one colorimetric parameter associated therewith to obtain filtered cosmetic compositions, wherein the at least one colorimetric parameter associated with each filtered cosmetic composition satisfies a criterion for at least one target colorimetric parameter; - forming several clusters of filtered cosmetic compositions, such that filtered cosmetic compositions from the same cluster comprise the same ingredients; - selecting clusters of filtered cosmetic compositions, called selected clusters; - determining a target cosmetic composition according to the filtered cosmetic compositions of the selected cluster; The present invention relates to a method, comprising:

[0128] By virtue of the steps of filtering the cosmetic compositions and forming clusters of the cosmetic compositions, it is possible to cluster together cosmetic compositions whose at least one colorimetric parameter is similar to at least one target colorimetric parameter and whose components are equal, thus making it possible to quickly determine one or more cosmetic compositions whose at least one colorimetric parameter corresponds to the target colorimetric parameter.

[0129] According to a particular embodiment, the determination method according to the second invention comprises one or several of the following characteristics, taken alone or in any technically possible combination: - each filtered cosmetic composition comprises, for each of its components, the amount of said component in the filtered cosmetic composition; The decision step of the processing phase is - calculating for each ingredient of the filtered cosmetic compositions of the selected cluster the average amount of said ingredient from said filtered cosmetic compositions; determining, for each of said ingredients, an average cosmetic composition comprising an amount equal to the calculated average value of each of the ingredients; forming a target cosmetic composition according to the average cosmetic composition; and - during the determining step of the processing phase, forming the target cosmetic composition includes calculating at least one colorimetric parameter associated with the average cosmetic composition; and optimizing the amount of each component of the average cosmetic composition such that the at least one colorimetric parameter associated with the average cosmetic composition approaches the at least one target colorimetric parameter; the target cosmetic composition being an average cosmetic composition resulting from optimizing the amounts of its ingredients; - during a selection step of the processing phase, M clusters of filtered cosmetic compositions are selected, where M is greater than or equal to 2, and during a determination step of the processing phase, a corresponding target cosmetic composition is determined for each selected cluster; - during a selection step of the processing phase, M clusters are selected such that, for each selected cluster, only ingredients contained in the selected cluster are included in the filtered cosmetic composition of the selected cluster; the method includes a training phase prior to the processing phase, the training phase comprising: receiving a set of feasibility constraints for the cosmetic composition; collecting a set of training data, each training data item being specific to a cosmetic composition that complies with the set of feasibility constraints received; Each training data item is A set of sizes representing the amount of each ingredient in the cosmetic composition; □ The value of at least one colorimetric parameter associated with the cosmetic composition; and training an artificial intelligence model based on a set of training data to obtain a trained model; Including, During the processing phase, during the step of obtaining at least one colorimetric parameter associated with each cosmetic composition under test, each colorimetric parameter is obtained by applying the trained model to a corresponding cosmetic composition under test; - the step of obtaining a set of cosmetic compositions under test of the processing phase, receiving a number N corresponding to the number of cosmetic compositions under test, the number N being greater than or equal to 2; generating N test data items, each test data item relates to a cosmetic composition whose ingredients are selected from a predefined list of ingredients, such that the cosmetic composition associated with each test data item is within a reduced feasible cosmetic composition space, the reduced space comprising only cosmetic compositions that meet a set of feasibility constraints; N test data items represent the reduced feasible cosmetic composition space; a set of cosmetic compositions under test is formed from the generated N test data items; and - during the processing phase, generating the N test data items of the obtaining step includes calculating, for each test data item, the magnitude of the N test data items such that the distance between said test data item and other test data items is maximal; each magnitude of each test data item represents the amount of the ingredient in the corresponding cosmetic composition being tested; - at least one colorimetric parameter is a triplet of values ​​characterizing the color of the hair after applying the cosmetic composition, or the color fading value of the cosmetic composition after washing the hair, or A selectivity value that characterizes the color difference between the root and tip of the hair. and - the treatment phase further comprises the step of preparing at least one sample of the target cosmetic composition for application on hair to verify the value of the at least one colorimetric parameter.

[0130] A second invention also relates to a computer program product having stored thereon a computer program comprising program instructions, the computer program being loaded onto a data processing unit and performing such a method when the computer program is implemented on the data processing unit.

[0131] A second invention also relates to an electronic device for determining a target cosmetic composition for dyeing hair, in particular hair, according to at least one target colorimetric parameter, wherein the electronic determination device comprises a processing unit capable of implementing such a method.

[0132] A second aspect of the invention also relates to a readable information medium on which a computer program product with program instructions is stored, the computer program being loaded onto a data processing unit and implementing such a determination method when the computer program is implemented on the data processing unit.

[0133] Other features and advantages of the second invention will become apparent on reading the following description of embodiments of the second invention, given purely by way of example and with reference to the drawings, in which:

[0134] In FIG. 5, an electronic device 1010 for determining a target cosmetic composition 1011* for dyeing hair 1012, in particular hair, according to at least one target colorimetric parameter 1013* is represented.

[0135] 5, target cosmetic composition 1011* is a cosmetic composition for dyeing hair 1012. Preferably, target cosmetic composition 1011* is capable of being applied to a user's hair 1012 to dye said hair 1012.

[0136] The target cosmetic composition 1011* comprises multiple ingredients 1014 that are capable of interacting with each other and with the user's hair 1012 to color it.

[0137] Preferably, during this interaction, ingredients 1014 of target cosmetic composition 1011* begin by bleaching hair 1012. Then ingredients 1014 of cosmetic composition 1011 interact with the bleached hair to fix a pigment of the selected color, thus forming dyed hair 1012*.

[0138] The dyed hair 1012* has a color that corresponds to at least one target colorimetric parameter 1013*.

[0139] Ingredient 1014 is present in target cosmetic composition 1011* in the form of, for example, a powder, gel, emulsion, or oil.

[0140] The electronic determination device 1010 is configured to use the at least one target colorimetric parameter 1013* to determine a set of magnitudes representing the amounts of each ingredient 1014 that form the target cosmetic composition 1011*, which, when applied to a user's hair 1012, is intended to produce a hair coloring that corresponds to at least the target colorimetric parameter 1013*.

[0141] Referring to FIG. 6, the decision device 1010 comprises a processing unit 1015 .

[0142] The decision device 1010 optionally further comprises a display screen 1016 and / or a unit 1017 for producing the cosmetic composition 1011 .

[0143] The processing unit 1015 may, for example, comprise a computing device that interacts with a computer program product. For example, the processing unit 1015 may be a computer.

[0144] The computer comprises, for example, a processor comprising a data processing unit, a memory, an information medium reader, and optionally a human-machine interface.

[0145] The computer program product includes an information medium.

[0146] The information medium is typically a computer-readable medium by a data processing unit. The readable data medium is a medium adapted to store electronic instructions and capable of being coupled to a computer system bus.

[0147] For example, the information medium is a USB flash disk, a floppy disk or flexible disk ("floppy disk"), an optical disk, a CD-ROM, a magneto-optical disk, a ROM memory, a RAM memory, an EPROM memory, an EEPROM memory, a magnetic card, or an optical card.

[0148] A computer program comprising program instructions is stored on an information medium.

[0149] The computer program may be loaded onto the processing unit 1015 and, when implemented on the processing unit of the computer, is adapted to perform a method for determining at least one target cosmetic composition 1011*, such a determination method being described herein below.

[0150] The operation of the electronic device 1010 implementing the method for determining a target cosmetic composition 1011* will now be described with reference to the flowchart of FIG. 7 and the examples of FIGS.

[0151] The determination method optionally includes a training phase 1100 .

[0152] The training phase 1100 includes receiving 1110 a set of feasibility constraints for the cosmetic composition 1011 .

[0153] Preferably, each cosmetic composition 1011 comprises at least one ingredient 1014 of a first type and at least one ingredient 1014 of a second type.

[0154] Each component 1014 of the first type is, for example, a base, and each component 1014 of the second type is, for example, a coupler.

[0155] Preferably, each component 1014 is selected from a predefined list of components.

[0156] Preferably, the set of feasibility constraints for the cosmetic composition 1011 includes the following constraints: - the ratio between the amount of the first type of component 1014 and the amount of the second type of component 1014 is between a first threshold value and a second threshold value; - the total amount of the cosmetic composition 1011 is less than the third threshold value; - the amount of each ingredient 1014 in the cosmetic composition 1011 is less than a fourth threshold; - the number of ingredients 1014 in the cosmetic composition 1011 is less than a fifth threshold Advantageously, the set of constraints comprises each of the above constraints.

[0157] These constraints are optionally derived from chemical principles that ensure the viability of the cosmetic composition 1011. The expression "viability of the cosmetic composition 1011" is understood herein to mean that a composition that meets the set of constraints is effective and / or does not pose a risk to the user and / or complies with product specifications and is preferably environmentally friendly.

[0158] The optional training phase 1100 further includes a step 1120 of collecting a set of training data 1018 .

[0159] Each training data item 1018 in the set of training data 1018 is specific to a cosmetic composition 1011 that complies with the set of feasibility constraints received during the receiving step 1110 .

[0160] Each training data item 1018 comprises a set of magnitudes representing the amount of each ingredient 1014 in the cosmetic composition 1011 and the value of at least one colorimetric parameter 1013 associated with the cosmetic composition 1011 .

[0161] Each magnitude is, for example, the amount of the corresponding ingredient 1014 in the associated cosmetic composition 1011 .

[0162] Each amount of ingredient 1014 is optionally a mass of the ingredient, a volume of the ingredient, a weight percentage of ingredient 1014 in cosmetic composition 1011, or a volume percentage of ingredient 1014 in cosmetic composition 1011.

[0163] The or each colorimetric parameter 1013 characterizes a cosmetic composition 1011 for coloring hair 1012 .

[0164] Preferably, each colorimetric parameter 1013 represents a visual effect on the hair 1012 to which the cosmetic composition 1011 is applied.

[0165] Advantageously, the at least one colorimetric parameter 1013 is: - a triplet of values ​​characterizing the color of the hair 1012 after applying the cosmetic composition 1011, and / or - the color fading value of the cosmetic composition 1011 after washing the hair 1012, and / or - a selectivity value characterizing the color difference between the root and tip of the hair 1012 Equipped with.

[0166] Preferably, each training data item 1018 comprises each of the above values.

[0167] The value triplet is, for example, a CIE L*a*b* color space value triplet.

[0168] The CIE L*a*b* color space, often abbreviated as CIELAB, is a color space for surface colors defined by the International Commission on Illumination (CIE) in 1976. It is based on the CIE XYZ system of evaluation, abandoning linearity to more accurately highlight differences between colors perceived by the human eye. In this model, three dimensions characterize a color: lightness (L*), derived from luminance (Y) in the XYZ evaluation; and two parameters, a* and b*, which describe the color difference from the color of a gray surface with the same lightness, as chrominance. The definition of gray, uncolored, and achromatic surfaces implies that the composition of the light illuminating a colored surface is explicitly stated. This light source is often a daylight source corresponding to the D65 normalization standard.

[0169] Also in this case, the triplet of values ​​comprises a lightness value L* and two color difference values ​​a*, b* from a gray surface color having the same lightness.

[0170] Any other color space may be envisaged.

[0171] Alternatively, the triplet of values ​​is an RGB triplet, which then comprises a value R for red, a value G for green and a value B for blue.

[0172] More generally, any representation that makes it possible to relate spectra can also be used in this context.

[0173] Hair color is not equal at all points on the hair, in particular the tips of the hair 1012 are generally lighter than the roots, so the selectivity value is an important parameter.

[0174] Preferably, during a collection step 1120 of the training phase 1100 , a set of training data 1018 is collected, at least in part, via a sensor capable of measuring at least one colorimetric parameter 1013 .

[0175] The sensor is, for example, a spectrocolorimeter.

[0176] According to one embodiment, the sensor is capable of capturing an image of a zone of hair of an individual and extracting colorimetric measurements from the captured image.

[0177] Preferably, each training data item 1018 is derived from a cosmetic composition 1011 that is feasible, ie, conforms to a set of feasibility constraints.

[0178] Preferably, each training data item corresponds to a cosmetic composition that is commercially available, and therefore, prior to marketing of said cosmetic composition, its colorimetric parameters 1013 have been measured during testing on hair 1012.

[0179] Further, optionally, the cosmetic compositions associated with the training data are pre-selected by an expert chemist from all commercially available cosmetic compositions.

[0180] The optional training phase 1100 further includes a step 1130 of training the artificial intelligence model using the set of training data 1018 to obtain a trained model.

[0181] Preferably, the artificial intelligence model comprises at least one of the following models: support vector machine, random forest, gradient boosting mechanism, or Kriging, also known as Gaussian process regression.

[0182] In each of the above cases, the artificial intelligence model comprises adjustable parameters, and during training step 1130, each adjustable parameter is adjusted such that, when a respective set of magnitudes of training data items 1018 is given as input, the trained model provides as output one or more colorimetric parameters 1013 that are substantially equal to the colorimetric parameters of the training data items 1018.

[0183] It is clear that the goal of the training phase 1100 described above is to obtain a trained model.

[0184] According to a variant not shown, the determination method does not include a training phase 1100 .

[0185] For example, the method may alternatively include receiving a trained model. According to another example, the method may alternatively include nothing.

[0186] The method further includes a processing phase 1200 .

[0187] The processing phase 1200 includes a step 1210 of collecting at least one target colorimetric parameter 1013*, which is provided by an operator of the determination device 1010, for example.

[0188] The processing phase 1200 further includes a step 1220 of obtaining a set of cosmetic compositions 1011 under test.

[0189] The set of cosmetic compositions under test 1011 is preferably formed from a plurality of test data items 1019 .

[0190] For each cosmetic composition 1011 under test, the test data item 1019 comprises a set of magnitudes representing the amount of each ingredient 1014 in the cosmetic composition 1011 under test.

[0191] Preferably, the ingredients 1014 of each cosmetic composition 1011 under test are selected from the predefined list of ingredients 1014 described above.

[0192] Each cosmetic composition under test 1011 conforms to a set of feasibility constraints received during the training phase 1100 .

[0193] If the method does not include a training phase 1100, the step 1220 of obtaining a set of cosmetic compositions under test 1011 further includes receiving said set of feasibility constraints.

[0194] The step 1220 of obtaining the set of cosmetic compositions under test 1011 advantageously includes receiving a number N corresponding to the number of cosmetic compositions under test 1011. Optionally, the number N is equal to or greater than 2. Preferably, the number N is greater than 100, for example, equal to 150.

[0195] The obtaining step 1220 further advantageously includes generating N test data items 1019, each test data item 1019 relating to a cosmetic composition 1011 whose ingredients 1014 are selected from a predefined list of ingredients 1014, such that the cosmetic composition 1011 associated with each test data item 1019 is within a reduced feasible cosmetic composition space 1020. The reduced feasible cosmetic composition space 1020 comprises only cosmetic compositions 1011 that comply with a set of feasibility constraints. The test data items 1019 represent the reduced feasible cosmetic composition space 1020.

[0196] The expression "test data items 1019 representing the reduced space 1020" is understood in this specification to mean that the test data items 1019 are appropriately selected in the reduced space 1020 so as to substantially cover the entire reduced space 1020.

[0197] 8 shows, via a two-dimensional block diagram, a set 1022 representing all possible cosmetic compositions 1011, and a reduced space 1020. It is clear that this set 1022 and this space 1020 cannot in fact be represented two-dimensionally, but rather can be represented in a number of dimensions equal to the number of ingredients 1014 in the predefined list of ingredients.

[0198] In Figure 8, a set 1022 of all possible cosmetic compositions 1011 corresponds to the map in Figure 8. As can be seen in Figure 8, a reduced space 1020 is included in said set 1022 because reduced space 1020 comprises only cosmetic compositions 1011 from set 1022 that meet the feasibility constraints.

[0199] 8, one can see the test data items 1019 in a reduced space 1020. For example, ten test data items 1019 are represented in FIG.

[0200] Preferably, generating the N test data items 1019 includes calculating, for each test data item 1019, the size of the N test data items 1019 such that the distance between said test data item 1019 and other test data items 1019 is maximum.

[0201] Each magnitude of each test data item 1019 represents the amount of ingredient 1014 in the corresponding cosmetic composition 1011 under test.

[0202] Generating N test data items 1019 includes calculating the magnitude among the N sets of magnitudes where the distance between each pair of the sets of magnitudes is greatest. Each magnitude in each set of magnitudes represents an amount of an ingredient 1014 in the cosmetic composition 1011 associated with that set of magnitudes, such that the cosmetic composition 1011 complies with a set of constraints. For example, test data item 1019 has a first magnitude equal to 4 moles of a first ingredient 1014A and 6 moles of a second ingredient 1014B.

[0203] Advantageously, each test data item 1019 is a vector comprising a coefficient for each component 1014 of a predefined list of components.

[0204] Each coefficient represents the amount of said ingredient 1014 in the respective cosmetic composition 1011 of the associated vector.

[0205] The distance between two test data items 1019, ie, between two vectors, is defined, for example, by an algebraic norm according to the following formula:

[0206]

number

[0207] where X and Y are two sets of size P, x i,i=1,...,P and y i,i=1,…,P are the respective sizes of the sets X and Y, √ is the square root function, Σ is the summation operator.

[0208] The N vectors are calculated, for example, by applying a design of experiments technique, also known as the DoE technique. Advantageously, the DoE technique implements an SFD (Space-Filling Design) algorithm, which makes it possible to ensure that the selected vectors represent the reduced space 1020, even when using a reduced number of vectors. Indeed, when the number of vectors is low, the assumption of the law of large numbers is not sufficiently fulfilled to ensure that a random distribution of vectors can represent the reduced space 1020.

[0209] In FIG. 8, it can be seen that the test data items 1019 are spaced apart from one another to maximize the distance between the test data items 1019.

[0210] Alternatively, the obtaining step 1220 only involves collecting one or more test data items 1019 selected by an operator of the decision device 1010. In this case, the operator selects one or more cosmetic compositions 1011 by selecting, for each cosmetic composition 1011, ingredients 1014 from a predefined list of ingredients and the amount of each of said ingredients 1014. When selecting, the operator ensures that each selected cosmetic composition 1011 complies with a set of constraints.

[0211] The set of cosmetic compositions under test is formed by N test data items 1019 .

[0212] The processing phase 1200 further includes a step 1230 of obtaining at least one colorimetric parameter 1013 associated with each cosmetic composition 1011 under test.

[0213] Preferably, the obtaining step 1230 includes applying the trained model to the test data items 1019 forming the obtained set of cosmetic compositions 1011 to determine at least one colorimetric parameter 1013 for the or each cosmetic composition 1011 under test.

[0214] Preferably, each test data item 1019 is sequentially fed to the trained model during the obtaining step 1230. The trained model then determines, for each test data item 1019, associated colorimetric parameters 1013.

[0215] Advantageously, during the obtaining step 1230 , the cosmetic composition 1011 and colorimetric parameters 1013 associated with each test data item are stored in one or more memories of the processing unit 1015 .

[0216] Alternatively, the obtaining step 1230 includes receiving, from an operator of the device 1010, at least one colorimetric parameter 1013 associated with each cosmetic composition 1011 of the set of cosmetic compositions 1011 under test.

[0217] In this case, said at least one colorimetric parameter 1013 is measured by, for example, an operator of the determination device 1010 using a sensor, for example as explained above for the test data item 1019 .

[0218] This is particularly advantageous if the method does not include a training phase 1100 or the receipt of a trained model.

[0219] This variant is particularly advantageously combined with a variant of step 1220 of obtaining a set of cosmetic compositions 1011 under test, whereby test data items 1019 associated with the cosmetic compositions 1011 under test are received from an operator.

[0220] The processing phase 1200 further comprises a filtering step 1235 during which the cosmetic compositions 1011 under test are filtered according to at least one colorimetric parameter 1013 associated with them. The filtering step 1235 makes it possible to obtain filtered cosmetic compositions 1030 for which at least one colorimetric parameter 1013 meets the criteria for at least one target colorimetric parameter 1013*.

[0221] Preferably, the criterion is that the relative difference between at least one colorimetric parameter 1013 associated with each cosmetic composition 1011 under test and at least one target colorimetric parameter 1013* is less than a sixth threshold. The sixth threshold is, for example, equal to 10%. The filtered cosmetic compositions 1030 are only those cosmetic compositions 1011 under test for which the relative difference is less than the sixth threshold. In other words, the filtered cosmetic compositions 1030 are those cosmetic compositions 1011 under test for which each colorimetric parameter 1013 deviates from the at least one target colorimetric parameter 1013* by at most the sixth threshold.

[0222] The processing phase 1200 further comprises a step 1240 of forming several clusters 1035 of the filtered cosmetic compositions 1030. For each cluster 1035, the filtered cosmetic compositions 1030 comprise the same ingredients 1014. The forming step 1240 comprises, for example, the application of an unsupervised learning algorithm, such as a k-means algorithm.

[0223] The filtered cosmetic compositions 1030 in each cluster 1035 are substantially similar. In other words, two filtered cosmetic compositions 1030 that comprise the same ingredients 1014 but in substantially different proportions will not be included in the same cluster 1035. For example, a filtered cosmetic composition 1030 that comprises 10% ingredient 1014A and 90% ingredient 1014B will not be included in the same cluster 1035 as a filtered cosmetic composition 1030 that comprises 80% ingredient 1014A and 20% ingredient 1014B.

[0224] 9, several clusters 1035A, 1035B, 1035C, 1035D are represented by ellipses, and the filtered cosmetic compositions 1030 of each cluster 1035A, 1035B, 1035C, 1035D are represented by crosses.

[0225] 9, a first cluster 1035A comprises a filtered cosmetic composition 1030 formed from ingredients 1014A and 1014B. A second cluster 1035B comprises a filtered cosmetic composition 1030 formed from ingredients 1014A, 1014B, and 1014C. A third cluster 1035C comprises a filtered cosmetic composition 1030 formed from ingredients 1014C and 1014D in a ratio substantially equal to 80 / 20. A fourth cluster 1035D comprises a filtered cosmetic composition 1030 formed from ingredients 1014C and 1014D in a ratio substantially equal to 40 / 60.

[0226] The processing phase 1200 further includes a step 1250 of selecting at least one cluster 1035 of the filtered cosmetic compositions 1030, referred to as a selected cluster 1035*. The selected cluster 1035* is, for example, selected randomly.

[0227] According to a variant, during a selection step 1250, M clusters 1035 are selected, M being greater than or equal to two and advantageously equal to four.

[0228] To this end, the M selected clusters 1035* are preferably selected such that, for each selected cluster 1035*, only the ingredients 1014 included in the filtered cosmetic composition 1030 of that selected cluster 1035* are included in that selected cluster 1035*. In other words, the ingredients 1014 included in the filtered cosmetic composition 1030 of that selected cluster 1035* are only found in that cluster 1035*. Stated differently, the ingredients 1014 included in the filtered cosmetic composition 1030 of that selected cluster 1035* are not included in the filtered cosmetic compositions 1030 of the other selected clusters 1035*.

[0229] In other words, the two selected clusters 1035* do not contain filtered cosmetic compositions 1030 having the same ingredients 1014.

[0230] 9, the first cluster 1035A and the second cluster 1035B therefore cannot belong to the M selected clusters 1035*. Similarly, the third cluster 1035C and the fourth cluster 1035D therefore cannot belong to the M selected clusters 1035*.

[0231] Thus, in the example in FIG. 9, the selected cluster 1035* may be, for example: - a first cluster 1035A and a third cluster 1035C; - a first cluster 1035A and a fourth cluster 1035D, a second cluster 1035B and a third cluster 1035C, or - a second cluster 1035B and a fourth cluster 1035D is.

[0232] Preferably, the selected cluster 1035* is a cluster 1035 that further comprises filtered cosmetic compositions 1030 for which each associated colorimetric parameter 1013 is closest to at least one target colorimetric parameter 1013*.

[0233] The processing phase 1200 further includes a step 1260 of determining a target cosmetic composition 1011* according to the filtered cosmetic compositions 1030 of the selected cluster 1035*.

[0234] Advantageously, when several clusters 1035 are selected, during a determination step 1260, a target cosmetic composition 1011* is determined for each selected cluster 1035*.

[0235] The determining step 1260 preferably includes calculating, for each component 1014 of the filtered cosmetic compositions 1030 of the selected cluster 1035*, an average value of the amount of said component 1014 from said filtered cosmetic compositions 1030. In the example in Figure 9, if the selected cluster 1035* is the first cluster 1035A, then said calculating is a calculation of an average value of the amount of component 1014A in each filtered cosmetic composition 1030 of the first cluster 1035A, and a calculation of an average value of the amount of component 1014B in each filtered cosmetic composition 1030 of the first cluster 1035A.

[0236] The determining step 1260 preferably further includes determining, for each of said ingredients 1014, an average cosmetic composition 1040 comprising an amount equal to the calculated average value of each of the ingredients 1014. In Figure 9, the average cosmetic compositions 1040A, 1040B, 1040C, 1040D of each cluster 1035A, 1035B, 1035C, 1035D are represented by squares.

[0237] It should be appreciated that the average cosmetic composition 1040 is then the centroid of the filtered cosmetic compositions 1030 of the selected cluster 1035*.

[0238] Advantageously, the determining step 1260 further comprises forming the target cosmetic composition 1011* according to the average cosmetic composition 1040.

[0239] For example, target cosmetic composition 1011* is average cosmetic composition 1040.

[0240] Alternatively, determining the target cosmetic composition 1011* according to the average cosmetic composition 1040 includes calculating at least one colorimetric parameter 1013 associated with the average cosmetic composition 1040, and optimizing the amount of each component 1014 of the average cosmetic composition 1040 so that the at least one colorimetric parameter 1013 associated with the average cosmetic composition 1040 approaches the at least one target colorimetric parameter 1013*.

[0241] In other words, an optimization algorithm is applied to the difference between at least one colorimetric parameter 1013 associated with the average cosmetic composition 1040 and at least one target colorimetric parameter 1013* to reduce this difference by varying the amount of each component 1014 of the average cosmetic composition 1040.

[0242] According to this variation, the target cosmetic composition 1011* is the average cosmetic composition 1040 that results from optimizing the amount of its ingredients 1014.

[0243] The calculation of the at least one colorimetric parameter 1013 associated with the average cosmetic composition 1040 is implemented, for example, by a trained model.

[0244] Further, optionally, when M clusters 1035 are selected, the determining step 1260 is repeated for each selected cluster 1035*. Thus, a target composition 1011* is determined for each selected cluster 1035*.

[0245] Additionally, optionally, processing phase 1200 further includes step 1270 of sending the or each target cosmetic composition 1011* to display screen 1016. Display screen 1016 then displays the or each target cosmetic composition 1011* directed to the operator of decision device 1010.

[0246] According to another optional addition, processing phase 1200 further comprises a step 1280 of producing at least one sample of the or each target cosmetic composition 1011*. The at least one sample is preferably produced by unit 1017 for producing cosmetic compositions 1011.

[0247] Alternatively, the manufacturing step 1280 includes receiving from the operator a selection of one or more target cosmetic compositions 1011* to be manufactured from the determined target cosmetic compositions 1011*. The manufacturing unit 1017 then manufactures only the samples of the target cosmetic compositions 1011* selected by the operator.

[0248] A sample of the manufactured target cosmetic composition 1011* is intended to be applied on hair 1012, preferably a tress of hair 1012, to verify colorimetric parameters.

[0249] Thus, following application of each sample on hair 1012, colorimetric parameters are measured using the same sensors used to form training data 1018.

[0250] Alternatively, the processing phase 1200 does not include the sending step 1270 and / or the manufacturing step 1280 .

[0251] Alternatively, each cosmetic composition 1011 may be applied to any type of hair 1012, not just hair, for example, to the user's eyelashes, eyebrows, or beard.

[0252] Processing phase 1200 is preferably repeated multiple times, forming an iteration, in which the at least one target colorimetric parameter 1013* collected during collection step 1210 is distinct from the at least one target colorimetric parameter 1013* collected for the preceding iteration.

[0253] Preferably, during each iteration, i.e., during each successive iteration following the first iteration, step 1220 of obtaining a set of cosmetic compositions, step 1230 of obtaining at least one colorimetric parameter 1013 associated with each cosmetic composition 1011 under test, and step 1240 of forming are not performed.

[0254] All the variations and optional additions described above can be combined with each other.

[0255] Using the determination method according to the second invention, it is possible to automatically obtain the target cosmetic composition 1011* using at least one target colorimetric parameter 1013*.

[0256] Next, a third aspect of the invention will be described with reference to FIGS.

[0257] The third invention relates to a method for determining a target cosmetic composition for dyeing hair, in particular hair, according to at least one target colorimetric parameter. The third invention also relates to a related computer program product and an electronic determination device.

[0258] A third invention relates to the field of cosmetic products, preferably cosmetic compositions for dyeing hair, especially hair.

[0259] More generally, a "cosmetic product" is a product as defined in Regulation EC 1223 / 2009 of the European Parliament and of the Council of 30 November 2009 on cosmetic products.

[0260] The goal of the cosmetics industry is to enhance the experience of its consumers. In particular, there is a growing trend to offer products that are increasingly adapted to the specific needs and characteristics of users. This trend is commonly referred to as "customization."

[0261] Customization of cosmetic products and services can concern any part of the human body, but is particularly important for exposed body parts such as the face (make-up or care products) and, where applicable, the hair or beard (e.g., coloring products).

[0262] In the field of cosmetic compositions for dyeing hair, and in particular hair, new cosmetic compositions are regularly developed in order to adapt to user demands.

[0263] These new cosmetic compositions are intended, in the first case, to achieve a new color or a new visual effect on the hair to which they are applied. In the second case, they are intended to achieve a previously known color, but comprise ingredients not previously used for this purpose. This second case is particularly important when some ingredients are difficult to source, involve high costs, or pose environmental risks.

[0264] However, obtaining new cosmetic compositions results from a very long and complicated process.

[0265] First, expert chemists develop cosmetic composition prototypes that are derived from the experience of expert chemists and are governed by several chemical principles.

[0266] These prototypes are then individually tested in the laboratory to ensure that the visual effects obtained meet the expectations of expert chemists.

[0267] Occasionally, and unpredictably, the visual effects associated with newly developed prototypes fall short of expectations.

[0268] It is then known to discard the prototype in order to test a new prototype that does not necessarily have a link with the first one.

[0269] In this case, the prototype will not be used if it does not meet expectations.

[0270] Therefore, there is a need for tools that allow profiting from prototypes that fail to meet expectations.

[0271] For this purpose, the third invention relates to a method for adapting an initial cosmetic composition intended to dye hair, in particular hair, so that the value of at least one colorimetric parameter of the adapted cosmetic composition corresponds to a target value, comprising: The method is performed by an electronic adaptive device and includes a processing phase, the processing phase comprising: - obtaining a target value for at least one colorimetric parameter; - obtaining data regarding an initial cosmetic composition, the data including a plurality of ingredients and, for each ingredient, an associated amount; - obtaining an initial value of at least one colorimetric parameter for the initial cosmetic composition, the initial value resulting from measurements after application of the initial cosmetic composition on hair, the initial value differing from the target value by a first deviation; - adapting the initial cosmetic composition to compensate for a first deviation between the initial value and the target value, wherein the adaptation comprises: Varying the amount of each ingredient in the initial cosmetic composition; assessing the effect of each variation in the value of at least one colorimetric parameter using an automated decision model, the automated decision model having been previously trained to determine the value of the at least one colorimetric parameter according to the cosmetic composition; Including, the applied cosmetic composition is obtained after application; The present invention relates to a method, comprising:

[0272] Thanks to the adaptation step, in particular via an automatic decision model, it is possible to obtain from the initial composition an adapted cosmetic composition whose value of at least one colorimetric parameter approaches a target value. It is then possible to take advantage of previous studies that resulted in prototypes whose at least one colorimetric parameter did not have a value equal to the target value.

[0273] According to a particular embodiment, the adaptation method according to the third invention comprises one or several of the following characteristics, taken alone or in any technically possible combination: - the adapted cosmetic composition has an adapted value for at least one colorimetric parameter, the adapted value resulting from a measurement after application of the adapted cosmetic composition on hair, the adapted cosmetic composition being such that a second deviation between the adapted value and the target value is less than the first deviation; - The processing phase is Using the automated decision model, calculating a matrix, called the Jacobian matrix, that represents the variation in the value of at least one colorimetric parameter of the initial cosmetic composition due to variation in the amount of each component of the initial cosmetic composition. further comprising during the evaluation of the influence of each variation in the value of at least one colorimetric parameter in an adaptation step of the processing phase, said influence is evaluated by calculating a Jacobian matrix and a value of a cost function that depends on each variation; - The cost function conforms to the following formula: FC=||Vi-Vc+JΔmp|| where FC is the cost function, Vi is the initial value of at least one colorimetric parameter; Vc is the target value, J is the Jacobian matrix, Δmp is a vector having component amount fluctuations; - during the adaptation step of the processing phase, the variation and evaluation are repeated several times by the optimization algorithm; - during an adaptation step of the processing phase, the applied optimization algorithm is an optimization algorithm constrained by a set of cosmetic composition feasibility constraints, the initial cosmetic composition and the adapted cosmetic composition comply with a set of feasibility constraints; the method includes a training phase prior to the processing phase, the training phase comprising: receiving a set of feasibility constraints for the cosmetic composition; collecting a set of training data, each training data item being specific to a cosmetic composition that complies with the set of feasibility constraints received; Each training data item is A set of sizes representing the amount of each ingredient in the cosmetic composition; □ The value of at least one colorimetric parameter associated with the cosmetic composition; and training an artificial intelligence model based on a set of training data to obtain a trained model capable of determining, from the cosmetic composition, the value of at least one colorimetric parameter of the cosmetic composition; Including, During the adaptation step of the processing phase, the model for determining at least one cosmetic composition colorimetric parameter is the model trained during the training step of the training phase; - the processing phase further comprises the step of producing a sample of the adapted cosmetic composition for measuring the adapted value of at least one colorimetric parameter of the adapted cosmetic composition; - the processing phase is repeated at least once to form at least two iterations; During the step of obtaining the or each iteration of the treatment phase, the initial cosmetic composition is the adapted cosmetic composition obtained in the preceding iteration, and the initial value results from a measurement after application of the adapted cosmetic composition from the preceding iteration on the hair.

[0274] A third aspect of the invention also relates to a computer program product having stored thereon a computer program comprising program instructions, the computer program being loaded onto a data processing unit and performing such a method when the computer program is implemented on the data processing unit.

[0275] The third invention also relates to an electronic device for adapting an initial cosmetic composition intended to dye hair, in particular hair, so that the value of at least one colorimetric parameter of the adapted cosmetic composition corresponds to a target value, the device comprising: It also relates to an electronic device, wherein the electronic adaptation device is capable of implementing such an adaptation method.

[0276] A third aspect of the invention also relates to a readable information medium on which a computer program product with program instructions is stored, the computer program being loaded onto a data processing unit and which performs such an adaptation method when the computer program is implemented on the data processing unit.

[0277] Other features and advantages of the third invention will become apparent on reading the following description of embodiments of the third invention, given purely by way of example and with reference to the drawings, in which:

[0278] In FIG. 10, the initial cosmetic composition 2011 is intended to dye hair 2012, particularly hair. i 20, an electronic device 2010 is depicted for adapting a cosmetic composition 2011* such that the value of at least one colorimetric parameter of the adapted cosmetic composition 2011* corresponds to a target value V. The verb "corresponds to" means that the value of the at least one colorimetric parameter approaches the target value V with respect to an initial value, as described below.

[0279] Early Cosmetic Composition 2011 i 10 because, in one embodiment, the adaptive device 2010 itself may dispense the initial cosmetic composition 2011, as described below. i This is because it is possible to determine

[0280] 10, the adapted cosmetic composition 2011* is a cosmetic composition for dyeing hair 2012. Preferably, the adapted cosmetic composition 2011* is capable of being applied to a user's hair 2012 to dye said hair 2012.

[0281] The adapted cosmetic composition 2011* comprises a plurality of ingredients 2014 that are capable of interacting with each other and with the user's hair 2012 to tint it.

[0282] Preferably, during this interaction, component 2014 of applied cosmetic composition 2011* begins by bleaching hair 2012. Then, component 2014 of applied cosmetic composition 2011* interacts with the bleached hair to fix a pigment of the selected color, thus forming dyed hair 2012*.

[0283] The dyed hair 2012* has a color that corresponds to the target value Vc of at least one colorimetric parameter.

[0284] Ingredient 2014 is present in the applied cosmetic composition 2011* in the form of, for example, a powder, a gel, an emulsion, or an oil.

[0285] Electronic adaptive devices 2010, early cosmetic compositions 2011 i 2014 in the hair 2012, thus forming an adapted cosmetic composition 2011*, which, when applied to a user's hair 2012, is intended to produce a hair coloring corresponding to at least the target colorimetric parameter Vc.

[0286] Referring to FIG. 11, the adaptive device 2010 includes a processing unit 2015 .

[0287] The adaptation device 2010 optionally further comprises a display screen 2016 and / or a unit 2017 for producing the cosmetic composition 2011 .

[0288] The processing unit 2015 may, for example, comprise a calculator that interacts with a computer program product. For example, the processing unit 2015 may be a computer.

[0289] The computer comprises, for example, a processor comprising a data processing unit, a memory, an information medium reader, and optionally a human-machine interface.

[0290] The computer program product includes an information medium.

[0291] The information medium is typically a computer-readable medium by a data processing unit. The readable data medium is a medium adapted to store electronic instructions and capable of being coupled to a computer system bus.

[0292] For example, the information medium is a USB flash disk, a floppy disk or flexible disk ("floppy disk"), an optical disk, a CD-ROM, a magneto-optical disk, a ROM memory, a RAM memory, an EPROM memory, an EEPROM memory, a magnetic card, or an optical card.

[0293] A computer program comprising program instructions is stored on an information medium.

[0294] The computer program may be loaded onto the processing unit 2015, and when the computer program is implemented on the processing unit of the computer, it may be used to process the initial cosmetic composition 2011 to obtain the adapted cosmetic composition 2011*. i Such adaptation methods are described herein below.

[0295] Next, the initial cosmetic composition 2011 i The operation of the electronic device 2010 implementing the method for adapting will now be described with reference to the flowcharts of FIGS. 12 and 13 and the examples of FIGS. 14-16.

[0296] The adaptive method optionally includes a training phase 2100 .

[0297] The training phase 2100 involves training a set of feasibility constraints E cv 2110 includes receiving the

[0298] Preferably, each cosmetic composition 2011 comprises at least one ingredient 2014 of a first type and at least one ingredient 2014 of a second type.

[0299] Each component 2014 of the first type is, for example, a base, and each component 2014 of the second type is, for example, a coupler.

[0300] Preferably, each component 2014 is selected from a predefined list of components.

[0301] Preferably, a set of feasibility constraints E of the cosmetic composition 2011 cv is subject to the following constraints: - the ratio between the amount of the first type of component 2014 and the amount of the second type of component 2014 is between a first threshold value and a second threshold value; - the total amount of cosmetic composition 2011 is less than the third threshold; - the amount of each ingredient 2014 in the cosmetic composition 2011 is less than the fourth threshold; and - The number of ingredients 2014 in the cosmetic composition 2011 is less than the fifth threshold The present invention is provided with one or more of the following:

[0302] Advantageously, the set of constraints E cv has each of the above constraints.

[0303] These constraints are optionally derived from chemical principles that ensure the viability of the cosmetic composition 2011. The expression "viability of the cosmetic composition 2011" is understood herein to mean that a composition that meets the set of constraints is effective and / or does not pose a risk to the user and / or complies with product specifications and is preferably environmentally friendly.

[0304] The optional training phase 2100 further includes a step 2120 of collecting a set of training data 2018.

[0305] Each training data item 2018 in the set of training data 2018 is unique to a cosmetic composition 2011 that complies with the set of feasibility constraints received during the receiving step 2110 .

[0306] Each training data item 2018 comprises a set of magnitudes representing the amount of each ingredient 2014 in the cosmetic composition 2011 and a value of at least one colorimetric parameter associated with the cosmetic composition 2011 .

[0307] Each magnitude is, for example, the amount of the corresponding ingredient 2014 in the associated cosmetic composition 2011 .

[0308] Each amount of ingredient 2014 is optionally a mass of the ingredient, a volume of the ingredient, a mass percentage of ingredient 2014 in cosmetic composition 2011, or a volume percentage of ingredient 2014 in cosmetic composition 2011.

[0309] The or each colorimetric parameter characterizes a cosmetic composition 2011 for coloring hair 2012 .

[0310] Preferably, each colorimetric parameter represents a visual effect on the hair 2012 to which the cosmetic composition 2011 is applied.

[0311] Advantageously, at least one colorimetric parameter is - a triplet of values ​​characterizing the color of the hair 2012 after applying the cosmetic composition 2011, and / or - the color fading value of the cosmetic composition 2011 after washing the hair 2012, and / or - Selectivity value characterizing the color difference between the root and tip of hair 2012 Equipped with.

[0312] Preferably, each training data item 2018 comprises each of the above values.

[0313] The value triplet is, for example, a CIE L*a*b* color space value triplet.

[0314] The CIE L*a*b* color space, often abbreviated as CIELAB, is a color space for surface colors defined by the International Commission on Illumination (CIE) in 1976. It is based on the CIE XYZ system of evaluation, abandoning linearity to more accurately highlight differences between colors perceived by the human eye. In this model, three dimensions characterize a color: lightness (L*), derived from luminance (Y) in the XYZ evaluation; and two parameters, a* and b*, which describe the color difference from the color of a gray surface with the same lightness, as chrominance. The definition of gray, uncolored, and achromatic surfaces implies that the composition of the light illuminating a colored surface is explicitly stated. This light source is often a daylight source corresponding to the D65 normalization standard.

[0315] Also in this case, the triplet of values ​​comprises a lightness value L* and two color difference values ​​a*, b* from a gray surface color having the same lightness.

[0316] Any other color space may be envisaged.

[0317] Alternatively, the triplet of values ​​is an RGB triplet, which then comprises a value R for red, a value G for green and a value B for blue.

[0318] More generally, any representation that makes it possible to relate spectra can also be used in this context.

[0319] In reality, hair color is not equal at all points on the hair, in particular the ends of the hair 2012 are generally lighter than the roots, so the selectivity value is an important parameter.

[0320] Preferably, during the collection step 2120 of the training phase 2100, the set of training data 2018 is collected, at least in part, via a sensor capable of measuring the value of at least one colorimetric parameter.

[0321] The sensor is, for example, a spectrocolorimeter.

[0322] According to one embodiment, the sensor is capable of capturing an image of a zone of hair of an individual and extracting colorimetric measurements from the captured image.

[0323] Preferably, each training data item 2018 is feasible, i.e., subject to a set of feasibility constraints E cv Derived from cosmetic composition 2011 that conforms to.

[0324] Preferably, each training data item corresponds to a commercially available cosmetic composition, such that prior to marketing of said cosmetic composition, the values ​​of its colorimetric parameters have been measured during testing on hair 2012.

[0325] Further, optionally, the cosmetic compositions associated with the training data are pre-selected by an expert chemist from all commercially available cosmetic compositions.

[0326] The optional training phase 2100 further includes a step 2130 of training an artificial intelligence model using a set of training data 2018 to obtain a trained model.

[0327] Preferably, the artificial intelligence model comprises at least one of the following models: support vector machine, random forest, gradient boosting mechanism, or Kriging, also known as Gaussian process regression.

[0328] In each of the above cases, the artificial intelligence model comprises adjustable parameters, and during training step 2130, each adjustable parameter is adjusted such that, when a respective set of magnitudes of training data items 2018 is given as input, the trained model provides as output a value of at least one colorimetric parameter that is substantially equal to the value of at least one colorimetric parameter of the training data items 2018.

[0329] It is clear that the purpose of the training phase 2100 described above is to obtain a trained model.

[0330] According to a variant not shown, the adaptation method does not include a training phase 2100 .

[0331] For example, the method may alternatively include receiving a trained model.

[0332] The method further includes a processing phase 2200 .

[0333] The processing phase 2200 includes a step 2210 of obtaining a target value Vc for at least one colorimetric parameter, which target value Vc of the at least one colorimetric parameter is provided, for example, by an operator of the adaptive device 2010.

[0334] The processing phase involves preparing an initial cosmetic composition 2011 that includes several ingredients 2014 and associated amounts of each ingredient 2014. i 2220. Hereinafter, the term "initial cosmetic composition 2011" is used. i "Getting" and "Initial Cosmetic Composition 2011 i The expressions "obtain data relating to" are equivalent.

[0335] Referring to FIG. 13, the initial cosmetic composition 2011 iThe step 2220 of obtaining includes the substep 2221 of obtaining a set of cosmetic compositions under test 2011.

[0336] The set of cosmetic compositions under test 2011 is preferably formed from several test data items 2019.

[0337] For each cosmetic composition 2011 under test, the test data item 2019 comprises a set of magnitudes representing the amount of each ingredient 2014 in the cosmetic composition 2011 under test.

[0338] Preferably, the ingredients 2014 of each cosmetic composition 2011 under test are selected from the predefined list of ingredients 2014 described above.

[0339] Each cosmetic composition under test 2011 is subjected to a set of feasibility constraints E received during the training phase 2100. cv Fits.

[0340] If the method does not include a training phase 2100, the substep 2221 of obtaining a set of cosmetic compositions under test 2011 further includes receiving said set of feasibility constraints.

[0341] Substep 2221 of obtaining a set of cosmetic compositions under test 2011 advantageously includes receiving a number N corresponding to the number of cosmetic compositions under test 2011. Optionally, number N is equal to or greater than 2. Preferably, number N is greater than 100, for example equal to 150.

[0342] The obtaining sub-step 2221 further advantageously includes generating N test data items 2019, each test data item 2019 relating to a cosmetic composition 2011 whose ingredients 2014 are selected from a predefined list of ingredients 2014, such that the cosmetic composition under test 2011 associated with each test data item 2019 is within a reduced feasible cosmetic composition space 2020. The reduced feasible cosmetic composition space 2020 is defined by a set of feasibility constraints E cv The test data items 2019 represent a reduced viable cosmetic composition space 2020.

[0343] The expression "test data items 2019 representing reduced space 2020" is understood in this specification to mean that the test data items 2019 are appropriately selected in reduced space 2020 so as to substantially cover the entire reduced space 2020.

[0344] 14 shows, via a two-dimensional block diagram, a set 2022 representing all possible cosmetic compositions 2011, and a reduced space 2020. It is clear that this set 2022 and this space 2020 cannot in fact be represented two-dimensionally, but rather can be represented in a number of dimensions equal to the number of ingredients 2014 in the predefined list of ingredients.

[0345] In Figure 14, a set 2022 of all possible cosmetic compositions 2011 corresponds to the map in Figure 14. As can be seen in Figure 14, a reduced space 2020 is included in said set 2022 because the reduced space 2020 is a set of feasibility constraints E cv This is because it only contains cosmetic compositions 2011 that comply with the above.

[0346] 14, one can see the test data items 2019 in a reduced space 2020. For example, ten test data items 2019 are represented in FIG.

[0347] Preferably, generating the N test data items 2019 includes calculating, for each test data item 2019, the magnitude of the N test data items 2019 such that the distance between said test data item 2019 and the other test data items 2019 is maximum.

[0348] Each magnitude of each test data item 2019 represents the amount of ingredient 2014 in the corresponding cosmetic composition 2011 being tested.

[0349] Generating N test data items 2019 includes calculating the magnitude among the N sets of magnitudes where the distance between each pair of the sets of magnitudes is greatest. Each magnitude in the respective set of magnitudes represents an amount of an ingredient 2014 in the cosmetic composition 2011 associated with that set of magnitudes, such that the cosmetic composition 2011 complies with a set of constraints. For example, test data item 2019 has a first magnitude equal to 4 moles of a first ingredient 2014A and 6 moles of a second ingredient 2014B.

[0350] Advantageously, each test data item 2019 is a vector comprising a coefficient for each component 2014 of a predefined list of components.

[0351] Each coefficient represents the amount of said ingredient 2014 in the respective cosmetic composition 2011 of the associated vector.

[0352] The distance between two test data items 2019, i.e., between two vectors, is defined, for example, by an algebraic norm according to the following formula:

[0353]

number

[0354] where X and Y are two sets of size P, x i,i=1,...,P and y i,i=1,…,Pare the respective sizes of the sets X and Y, √ is the square root function, Σ is the summation operator.

[0355] The N vectors are calculated, for example, by applying a design of experiments technique, also known as a DoE technique. Advantageously, the DoE technique implements an SFD (Space-Filling Design) algorithm, which makes it possible to ensure that the selected vectors represent the reduced space 2020, even when using a reduced number of vectors. Indeed, when the number of vectors is low, the assumption of the law of large numbers is not sufficiently fulfilled to ensure that a random distribution of vectors can represent the reduced space 2020.

[0356] In FIG. 14, it can be seen that the test data items 2019 are spaced apart from one another to maximize the distance between the test data items 2019.

[0357] Alternatively, obtaining substep 2221 only involves collecting one or more test data items 2019 selected by an operator of adaptive device 2010. In this case, the operator selects one or more cosmetic compositions 2011 by selecting, for each cosmetic composition 2011, ingredients 2014 from a predefined list of ingredients and the amount of each of said ingredients 2014. When selecting, the operator ensures that each selected cosmetic composition 2011 complies with a set of constraints.

[0358] The set of cosmetic compositions under test is formed by N test data items 2019.

[0359] The step 2220 of obtaining an initial cosmetic composition further includes the sub-step 2222 of obtaining a value for at least one colorimetric parameter associated with each cosmetic composition 2011 under test.

[0360] Preferably, obtaining substep 2222 includes applying the trained model to test data items 2019 forming the obtained set of cosmetic compositions 2011 to determine a value of at least one colorimetric parameter for the or each cosmetic composition 2011 under test.

[0361] Preferably, during obtaining substep 2222, each test data item 2019 is sequentially fed to the trained model. The trained model then determines, for each test data item 2019, the values ​​of the associated colorimetric parameters.

[0362] Advantageously, during the obtaining sub-step 2222, the cosmetic composition 2011 and the value of at least one colorimetric parameter associated with each test data item are stored in one or more memories of the processing unit 2015.

[0363] Alternatively, obtaining substep 2222 includes receiving, from an operator of said device 2010, a value of at least one colorimetric parameter associated with each cosmetic composition 2011 of the set of cosmetic compositions 2011 under test.

[0364] In this case, the value of at least one colorimetric parameter is measured, for example, by an operator of the adaptive device 2010 using a sensor, for example, as described above for test data item 2019.

[0365] This variant is particularly advantageously combined with a variant of substep 2221 of obtaining a set of cosmetic compositions 2011 under test, whereby test data items 2019 associated with the cosmetic compositions 2011 under test are received from an operator.

[0366] The step 2220 of obtaining a cosmetic composition further comprises a filtering substep 2223, during which the cosmetic compositions 2011 under test are filtered according to the value of at least one colorimetric parameter associated with them. The filtering substep 2223 makes it possible to obtain filtered cosmetic compositions 2030 for which the value of at least one colorimetric parameter meets a criterion with respect to the target value Vc.

[0367] Preferably, the criterion is that the relative difference between the value of at least one colorimetric parameter associated with each cosmetic composition 2011 under test and the target value Vc is less than a sixth threshold. The filtered cosmetic compositions 2030 are only those cosmetic compositions 2011 under test for which said relative difference is less than the sixth threshold. In other words, the filtered cosmetic compositions 2030 are those cosmetic compositions 2011 under test for which the value of each colorimetric parameter deviates from the target value Vc by at most the sixth threshold.

[0368] Early Cosmetic Composition 2011 i The step 2220 of obtaining further comprises a sub-step 2224 of forming several clusters 2035 of filtered cosmetic compositions 2030. For each cluster 2035, the filtered cosmetic compositions 2030 comprise the same ingredients 2014. The forming sub-step 2224 comprises, for example, the application of an unsupervised learning algorithm, such as a k-means algorithm.

[0369] The filtered cosmetic compositions 2030 in each cluster 2035 are substantially similar. In other words, two filtered cosmetic compositions 2030 that comprise the same ingredients 2014, but in substantially different proportions, will not be included in the same cluster 2035. For example, a filtered cosmetic composition 2030 that comprises 10% ingredient 2014A and 90% ingredient 2014B will not be included in the same cluster 2035 as a filtered cosmetic composition 2030 that comprises 80% ingredient 2014A and 20% ingredient 2014B.

[0370] In Figure 15, several clusters 2035A, 2035B, 2035C, 2035D are represented by ellipses, and the filtered cosmetic compositions 2030 of each cluster 2035A, 2035B, 2035C, 2035D are represented by crosses.

[0371] As seen in Figure 15, first cluster 2035A comprises filtered cosmetic composition 2030 formed from ingredients 2014A and 2014B. Second cluster 2035B comprises filtered cosmetic composition 2030 formed from ingredients 2014A, 2014B, and 2014C. Third cluster 2035C comprises filtered cosmetic composition 2030 formed from ingredients 2014C and 2014D in a ratio substantially equal to 80 / 20. Fourth cluster 2035D comprises filtered cosmetic composition 2030 formed from ingredients 2014C and 2014D in a ratio substantially equal to 40 / 60.

[0372] Early Cosmetic Composition 2011 i The step 2220 of obtaining further comprises a substep 2225 of selecting at least one cluster 2035 of the filtered cosmetic compositions 2030, referred to as a selected cluster 2035*. The selected cluster 2035* is, for example, selected randomly.

[0373] According to a variant, during a selection substep 2225, M clusters 2035 are selected, M being greater than or equal to two and advantageously equal to four.

[0374] To this end, the M selected clusters 2035* are preferably selected such that, for each selected cluster 2035*, only the ingredients 2014 included in the filtered cosmetic composition 2030 of that selected cluster 2035* are included in that selected cluster 2035*. In other words, the ingredients 2014 included in the filtered cosmetic composition 2030 of that selected cluster 2035* are exclusive to that cluster 2035*. Stated differently, the ingredients 2014 included in the filtered cosmetic composition 2030 of that selected cluster 2035* are not included in the filtered cosmetic compositions 2030 of the other selected clusters 2035*.

[0375] In other words, the two selected clusters 2035* do not contain filtered cosmetic compositions 2030 that have the same ingredients 2014.

[0376] 14, the first cluster 2035A and the second cluster 2035B therefore cannot belong to the M selected clusters 2035*. Similarly, the third cluster 2035C and the fourth cluster 2035D therefore cannot belong to the M selected clusters 2035*.

[0377] Thus, in the example in FIG. 15, the selected cluster 2035* may be, for example: - a first cluster 2035A and a third cluster 2035C; - a first cluster 2035A and a fourth cluster 2035D, - a second cluster 2035B and a third cluster 2035C, or - Second cluster 2035B and fourth cluster 2035D is.

[0378] Preferably, the selected cluster 2035* is the cluster 2035 that further comprises the filtered cosmetic compositions 2030 for which the value of each associated colorimetric parameter is closest to the target value Vc.

[0379] The step 2220 of obtaining initial cosmetic compositions includes filtering potential cosmetic compositions 2011 according to the filtered cosmetic compositions 2030 of the selected cluster 2035*. p The method further includes a sub-step 2226 of determining:

[0380] Advantageously, when several clusters 2035 are selected, during the determining substep 2226, the potential cosmetic compositions 2011 p is determined for each selected cluster 2035*.

[0381] Determining sub-step 2226 preferably includes calculating, for each component 2014 of the filtered cosmetic compositions 2030 of the selected cluster 2035*, the average value of the amount of said component 2014 from said filtered cosmetic compositions 2030. In the example in Figure 15, if the selected cluster 2035* is the first cluster 2035A, said calculating would be the calculation of the average value of the amount of component 2014A in each filtered cosmetic composition 2030 of the first cluster 2035A, and the calculation of the average value of the amount of component 2014B in each filtered cosmetic composition 2030 of the first cluster 2035A.

[0382] Determining sub-step 2226 preferably further includes determining, for each of said ingredients 2014, an average cosmetic composition 2040 comprising an amount equal to the calculated average value of each of the ingredients 2014. In Figure 15, the average cosmetic compositions 2040A, 2040B, 2040C, 2040D of each cluster 2035A, 2035B, 2035C, 2035D are represented by squares.

[0383] It should be appreciated that the average cosmetic composition 2040 is then the centroid of the filtered cosmetic compositions 2030 of the selected cluster 2035*.

[0384] Advantageously, the determining substep 2226 determines the potential cosmetic compositions 2011 according to the average cosmetic composition 2040. p forming a

[0385] For example, Potential Cosmetic Compositions 2011 p is the average cosmetic composition 2040.

[0386] Alternatively, the potential cosmetic composition 2011 may be calculated according to the average cosmetic composition 2040. p Determining includes calculating a value of at least one colorimetric parameter associated with the average cosmetic composition 2040, and optimizing the amount of each ingredient 2014 of the average cosmetic composition 2040 such that the value of the at least one colorimetric parameter associated with the average cosmetic composition 2040 approaches the target value Vc.

[0387] In other words, a first optimization algorithm is applied to the difference between the value of at least one colorimetric parameter associated with the average cosmetic composition 2040 and the target value Vc to reduce this difference by varying the amount of each component 2014 of the average cosmetic composition 2040.

[0388] According to this variant, the potential cosmetic composition 2011 p is the average cosmetic composition 2040 that results from optimizing the amount of that ingredient 2014.

[0389] The calculation of the at least one colorimetric parameter associated with the average cosmetic composition 2040 is implemented, for example, by a trained model.

[0390] Further, optionally, when M clusters 2035 are selected, the determining substep 2226 is repeated for each selected cluster 2035*. p is determined for each selected cluster 2035*.

[0391] Additionally, optionally, the step of obtaining an initial cosmetic composition 2220 may include displaying on the display screen 2016 the or each potential cosmetic composition 2011. p The display screen 2016 then displays the or each potential cosmetic composition 2011 directed to the operator of the adaptive device 2010. p Display.

[0392] According to another optional addition, the obtaining step 2220 may include obtaining the or each potential cosmetic composition 2011. p The at least one sample is preferably produced by the unit 2017 for producing the cosmetic composition 2011.

[0393] A sample of the or each potential cosmetic composition is then prepared. p The initial cosmetic composition 2011 i is.

[0394] Advantageously, the manufacturing sub-step 2228 comprises: p Potential cosmetic compositions to be produced from 2011 p The manufacturing unit 2017 then receives from the operator a selection of the potential cosmetic composition 2011 selected by the operator. p Only samples of this product will be produced.

[0395] In this case, the sample is prepared from the potential cosmetic composition 2011 p Only the initial cosmetic composition 2011 i is.

[0396] Next, the initial cosmetic composition 2011 i It should be understood that at least one colorimetric parameter of has a value substantially equal to the target value Vc according to the trained model.

[0397] Early Cosmetic Composition 2011 i is intended to be applied on hair 2012, preferably on a tress of hair 2012, in order to verify the value of at least one colorimetric parameter.

[0398] Therefore, the initial cosmetic composition 2011 in hair 2012 i Following application of the at least one colorimetric parameter, a value of the at least one colorimetric parameter is measured using the same sensor used to form the training data 2018 to form an initial value Vi of the at least one colorimetric parameter.

[0399] As described above, the application device 2010 then applies the initial cosmetic composition 2011 i Do not receive the initial cosmetic composition 2011 by itself i Please understand that we will prepare the following.

[0400] Alternatively, the initial cosmetic composition 2011 i The step 2220 of obtaining does not include any of the sub-steps described above. Instead, during the obtaining step 2220, the adaptation device 2010 receives an initial cosmetic composition 2011 from an operator of the adaptation device 2010. i Receive.

[0401] In this case, the operator will select the initial cosmetic composition 2011 i is the set of feasibility constraints E cv The operator ensures that the initial cosmetic composition 2011 i22. Preferably, in this case, the order of step 2210 of obtaining the target value and step 2220 of obtaining the initial cosmetic composition is reversed, and the target value Vc is determined to be the value of the initial cosmetic composition 2011. i is determined by applying a trained model to

[0402] According to this variant, the initial cosmetic composition 2011 i However, the data is nevertheless applied to the hair 2012 and the value of at least one associated colorimetric parameter is measured using the same sensor used to form the training data 2018 to form an initial value Vi of the at least one colorimetric parameter.

[0403] In some cases, the trained model is not perfect and the initial values ​​Vi resulting from measurements do not correspond perfectly to the target values ​​Vc.

[0404] The processing phase 2200 includes a step 2230 of obtaining an initial value Vi, where the initial value Vi is the initial cosmetic composition 2011 in the hair 2012. i The initial value Vi differs from the target value by a first deviation.

[0405] Optionally, the processing phase 2200 includes a step 2240 of calculating a matrix, called the Jacobian matrix J, using an automatic determination model of the values ​​of at least one colorimetric parameter using the cosmetic composition. The automatic determination model is preferably a trained model. The Jacobian matrix J is calculated based on the initial cosmetic composition 2011. i The initial cosmetic composition 2011, with variations in the amount of each ingredient 2014 i represents the variation of the value V of at least one colorimetric parameter of

[0406] The Jacobian matrix J is preferably the initial cosmetic composition 2011 iis calculated for the operating point of the trained model, which is equal to

[0407] Optionally, the relative amplitudes amp for calculating the Jacobian matrix J are predefined. i For each component 2014, a pair of calculated cosmetic compositions 2045A, 2045B, 2046A, 2046B is calculated. For each pair, the first calculated cosmetic composition 2045A, 2046A is calculated based on the initial cosmetic composition 2011. i The amount of each component 2014 is the same as that of the initial cosmetic composition 2011, but the amount is different from the relative amplitude and i In other words, the amount of said component 2014 in the first calculated composition 2045A, 2046A conforms to the following formula:

[0408]

number

[0409] However, mp i is the amount of the i-th ingredient in the first calculated cosmetic composition 2045A, 2046A;

[0410]

number

[0411] The initial composition 2011 i is the amount of the i-th component in

[0412] For the same pair, the second calculated cosmetic compositions 2045B and 2046B are calculated based on the initial cosmetic composition 2011. i The amount of each component 2014 is the same as that of the initial cosmetic composition 2011, but the amount is different from the relative amplitude and i In other words, the amounts of the components in the second calculated compositions 2045B, 2046B comply with the following formula:

[0413]

number

[0414] However, mp i is the amount of the i-th ingredient in the second calculated cosmetic composition.

[0415] In Figure 16, this calculation is based on the initial cosmetic composition 2011 i is shown in a simple example with only two components 2014. Figure 16 represents a frame of reference where the quantity of the first component is shown on the x-axis and the quantity of the second component is shown on the y-axis.

[0416] In Figure 16, the initial cosmetic composition 2011 i are represented by circles. A first calculated cosmetic composition 2045A and a second calculated cosmetic composition 2045B associated with a first ingredient are represented by crosses. A first calculated cosmetic composition 2046A and a second calculated cosmetic composition 2046B associated with a second ingredient are represented by squares.

[0417] In FIG. 16, for each of the first calculated cosmetic composition 2045A and the second calculated cosmetic composition 2045B associated with the first component, the amount of the second component is calculated based on the initial composition 2011. i Note that the amount of the second component in the calculated cosmetic composition 2046A is equal to the amount of the second component in the calculated cosmetic composition 2046B. Similarly, for each of the first calculated cosmetic composition 2046A and the second calculated cosmetic composition 2046B associated with a second component, the amount of the first component is equal to the amount of the first component in the initial composition.

[0418] For each ingredient 2014, a first calculated value VA of the at least one colorimetric parameter is determined by applying the automated determination model to the first calculated cosmetic composition 2045A, 2046A, and a second calculated value VB of the at least one colorimetric parameter is determined by applying the automated determination model to the second calculated cosmetic composition 2045B, 2046B.i The vector called ���� is determined by calculating the average of the first calculated value VA and the second calculated value VB weighted by the relative amplitude amp. In other words, the gradient vector grad i is fitted to the following equation:

[0419]

number

[0420] However, grad i The initial cosmetic composition 2011 i is the gradient associated with the i-th component 2014 of

[0421] Calculation of cosmetic composition pairs 2045A, 2045B, 2046A, 2046B, calculation of values ​​VA, VB, and gradient vector grad i Calculation of initial cosmetic composition 2011 i This is repeated for each component 2014 of i will be obtained.

[0422] Then, these gradient vectors grad i are arranged in a matrix forming the Jacobian matrix J, for example according to the following formula: [Number 5] J=[grad1... grad i ...grad n ] However, n is the initial cosmetic composition 2011 i is the number of components in 2014.

[0423] Additionally, the amount of each of its components 2014 is of magnitude Δmp i Only early cosmetic compositions 2011 iFor any cosmetic composition 2011 different from the above, the value V of at least one associated colorimetric parameter is represented by a first order linearization of the automated determination model using the following equation: [Number 6] V=Vi+JΔmp However, Vi is the initial value, Δmp is the magnitude of fluctuation in the amount of each component. i is a vector comprising:

[0424] The processing phase 2200 adjusts the initial cosmetic composition 2011 to compensate for a first deviation between the initial value Vi and the target value Vc. i The method further includes a step 2250 of adapting

[0425] The adapting step 2250 may, for example, adapt the initial cosmetic composition 2011 to form a varied cosmetic composition in which the amount of each ingredient conforms to the following formula: i This involves varying the amount of each component 2014 in the mixture. [Number 7] mp v =mp+Δmp However, mp v is a vector comprising the amount of each ingredient 2014 of the varied cosmetic composition; mp is the initial cosmetic composition 2011 i is a vector with the amount of each component of Δmp is a vector comprising, for each component 2014, the variation of the amount of this component 2014.

[0426] An adaptation step 2250 then involves using an automated decision model to evaluate the effect of each variation in the value of at least one colorimetric parameter.

[0427] Preferably, this influence is evaluated by calculating the value of a cost function FC, which depends on the initial values ​​Vi, the target values ​​Vc, the Jacobian matrix J, and optionally on said variations.

[0428] Preferably, the cost function FC conforms to the following equation: [Number 8] FC=||Vi-Vc+JΔmp|| where ||.|| is a mathematical norm.

[0429] It should be understood that the first deviation is then compensated when the value of the cost function is equal to 0. However, this value of the cost function FC is not always reachable.

[0430] Preferably, during the adaptation step 2250, the variation and evaluation are repeated several times by an optimization algorithm aimed at minimizing the cost function value F. In each iteration, the optimization algorithm determines the variation of the amount of each component 2014 according to the cost function value F of the previous iteration. Preferably, the variation is determined using the Jacobian matrix J. In practice, the initial values ​​V and the target values ​​V are constants for the optimization algorithm.

[0431] The problem to be solved by the optimization algorithm can then be formulated in this way.

[0432]

number

[0433] According to a variant, the optimization algorithm applied is based on a set of cosmetic composition feasibility constraints E cv The optimization problem to be solved by the optimization algorithm can then be formulated as follows:

[0434]

number

[0435] However, E cv is the set of feasibility constraints.

[0436] According to a variant, the cost function F C represents the variation in the amount of each component 2014 and is penalized by a magnitude that depends on a predefined coefficient α. In other words, the cost function F C fits, for example, the following equation: [Number 11] FC=||Vi-Vc+JΔmp||++α||Δmp|| 2

[0437] By penalizing the cost function FC, the cosmetic compositions obtained for the following iterations are compared with the initial composition 2011. i The above considerations for the optimization problems of Equation 9 and Equation 10 also apply when the cost function F C is fitted to Equation 11 instead of Equation 8.

[0438] The solution to the optimization problem is the initial cosmetic composition 2011 i provides an optimized variation Δmp* in the amount of each component 2014 of

[0439] Following the adapting step 2250, an adapted cosmetic composition 2011* is obtained. The adapted cosmetic composition 2011* is obtained by adjusting the initial cosmetic composition 2011 i The amount of each component 2014 of the adapted cosmetic composition 2011* is adjusted to the same amount as the initial composition 2011 by the corresponding optimized variation. i The amount of component 2014 in the formula (I) is different from the amount of component 2014 in the formula (I).

[0440] When the optimization problem solved by the optimization algorithm is the problem of Equation 10, the initial cosmetic composition 2011 i and the adapted cosmetic composition 2011* is subject to a set of feasibility constraints E cv Fits.

[0441] Further, optionally, processing phase 2200 further includes step 2260 of sending adapted cosmetic composition 2011* to display screen 2016. Display screen 2016 then displays adapted cosmetic composition 2011* directed to the operator.

[0442] According to another optional addition, the processing phase 2200 further comprises a step 2270 of producing at least one sample of the adapted cosmetic composition 2011*. The at least one sample is preferably produced by a unit 2017 for producing a cosmetic composition.

[0443] A sample of the adapted cosmetic composition 2011* is intended to be applied on hair 2012, preferably a tress of hair 2012, in order to verify the value of at least one colorimetric parameter.

[0444] Thus, following application of the sample on the hair 2012, an adapted value Va of at least one colorimetric parameter of the adapted cosmetic composition 2011* is measured using the same sensor used to form the training data 2018.

[0445] The adapted cosmetic composition 2011* is such that the second deviation between the adapted value Va and the target value Vc is less than the first deviation.

[0446] Preferably, processing phase 2200 is repeated at least once to form at least two iterations. In each iteration, initial cosmetic composition 2011 i During the step 2220 of obtaining the initial cosmetic composition 2011 i is the adapted composition 2011* of the previous processing phase 2200. Therefore, none of the sub-steps 2221, 2222, 2223, 2224, 2225, 2226, 2227, 2228 are performed again.

[0447] At each iteration, step 2210 of obtaining a target value is preferably not performed.

[0448] In each iteration, during a step 2230 of obtaining an initial value V of at least one colorimetric parameter, said initial value V is the adapted value V of the preceding processing phase 2200. In each iteration, during an optional step 2240 of calculating a Jacobian matrix J, the operating point for which the Jacobian matrix J is calculated is the adapted cosmetic composition 2011* of the preceding processing phase 2200.

[0449] Thus, an iteration-specific adapted value Va is measured for each iteration of the processing phase 2200. Preferably, the adapted values ​​form a sequence that converges towards a target value Vc.

[0450] Additionally, following an iteration of processing phase 2200, the obtained adapted cosmetic composition 2011* has an adapted value Va of at least one colorimetric parameter that is equal to the target value Vc.

[0451] Alternatively, each cosmetic composition can be applied to any type of hair 2012, not just hair, for example, to the user's eyelashes, eyebrows, or beard.

[0452] The above-described variations and additions may be combined in any technically possible combination. [Explanation of symbols]

[0453] 10 Electronic devices, electronic decision-making devices, decision-making devices 11, 1011, 2011 Cosmetic compositions 12, 1012, 2012 Hair, hair 12*, 1012*, 2012* Dyed hair 13 users 14, 1014, 1014C, 1014D, 2014, 2014C, 2014D components 14A First Component 14B Second component 15 Processing unit, data processing unit 16, 1016, 2016 display screen 17 Unit for producing cosmetic composition 11, production unit 18, 1018, 2018 training data, training data items 19, 1019, 2019 test data items 20, 1020, 2020 Reduced viable cosmetic composition space, reduced space, space 22, 1022, 2022 sets 1010 Electronic devices, electronic decision devices, decision devices, devices 1011* Target cosmetic composition, target composition 1013 Colorimetric Parameters 1013* target colorimetric parameters 1014A, 2014A 1st component, component 1014B, 2014B second component, component 1015, 2015 processing unit 1017 Units for producing cosmetic compositions 1011, production units 1030, 2030 Filtered cosmetic composition 1035, 2035 clusters 1035A, 2035A Cluster, First Cluster 1035B, 2035B cluster, second cluster 1035C, 2035C cluster, third cluster 1035D, 2035D cluster, 4th cluster 1035*, 2035* selected cluster, cluster 1040, 1040A, 1040B, 1040C, 1040D, 2040, 2040A, 2040B, 2040C, 2040D Average cosmetic composition 2010 Electronic Devices, Adaptive Devices, Electronic Adaptive Devices, Devices 2011 i Initial cosmetic composition, initial composition 2011* Adapted cosmetic composition, adapted composition 2011 p Potential cosmetic composition, potential composition 2017 Unit for producing cosmetic compositions 2011, production unit, unit for producing cosmetic compositions 2045A, 2046A Calculated cosmetic composition, first calculated cosmetic composition, first calculated composition 2045B, 2046B Calculated cosmetic composition, second calculated cosmetic composition, second calculated composition

Claims

1. A method for determining at least one colorimetric parameter characterizing hair (12), in particular a cosmetic composition (11) for dyeing hair, said method being implemented by an electronic determination device (10) and comprising a training phase (100), said training phase (100) comprising: - receiving (110) a set of feasibility constraints for a cosmetic composition; - collecting (120) a set of training data (18), each training data item (18) being specific to a cosmetic composition (11) that complies with said set of received feasibility constraints; Each training data item (18) a set of sizes representing the amount of each component (14) of said cosmetic composition (11); the value of said at least one colorimetric parameter associated with said cosmetic composition (11); and and - training (130) an artificial intelligence model based on said set of training data (18) to obtain a trained model; Including, The method further comprises a processing phase (200), wherein the processing phase (200) comprises: - obtaining (210) test data items (19) relating to a cosmetic composition (11) under test, wherein said cosmetic composition (11) under test complies with said set of feasibility constraints received during said training phase (100); the test data items (19) comprising a set of magnitudes representing the amounts of each ingredient (14) of the cosmetic composition (11) under test; - applying (220) the trained model to the acquired test data items (19) to determine the value of the at least one colorimetric parameter of the cosmetic composition (11) under test; A method comprising:

2. 10. The method of claim 1, wherein each ingredient (14) of the test data item (19) for the cosmetic composition (11) under test is selected from a predefined list of ingredients.

3. During the obtaining step (210) of the processing phase (200), several test data items (19) are obtained, each test data item (19) being specific to a respective cosmetic composition (11) under test; During the applying step (220) of the processing phase (200), the trained model is applied to each acquired test data item to determine a value of the at least one colorimetric parameter associated with the test data item (19); The acquiring step (210) of the processing phase (200) - receiving a number N corresponding to the number of cosmetic compositions (11) under test, said number N being greater than or equal to 2; - generating N test data items (19), each test data item (19) relates to a cosmetic composition (11) whose ingredients (14) are selected from a predefined list of ingredients (14), such that the cosmetic composition (11) associated with each test data item (19) is within a reduced feasible cosmetic composition space (20); the reduced feasible cosmetic composition space (20) comprises only cosmetic compositions (11) that comply with the set of feasibility constraints; said N test data items (19) representing said reduced feasible cosmetic composition space (20); 3. The method of claim 2, comprising:

4. During the obtaining step of the processing phase (200), the step of generating N test data items (19) in the obtaining step (210) includes a step of calculating, for each test data item (19), a size of the N test data items (19) such that a distance between the test data item (19) and other test data items (19) is maximum; 4. The method of claim 3, wherein each magnitude of each test data item (19) represents the amount of ingredient (14) in the corresponding cosmetic composition (11) under test.

5. the predefined list of components comprises one or more components (14) of a first type and one or more components (14) of a second type; the or each component (14) of the first type is a base; 5. The method of any one of claims 2 to 4, wherein the or each component (14) of the second type is a coupler.

6. The set of feasibility constraints for the cosmetic composition (11) includes the following constraints: - the ratio between the amount of the first type of component (14) and the amount of the second type of component (14) is between a first threshold value and a second threshold value; - the total amount of said cosmetic composition (11) is less than a third threshold value; - the amount of each component (14) in said cosmetic composition (11) is less than a fourth threshold value; - the number of ingredients (14) in said cosmetic composition (11) is less than a fifth threshold value; The method of claim 5, comprising one or more of:

7. The at least one colorimetric parameter is - a triplet of values ​​(L*, a*, b*, R, G, B) characterizing the color of the hair (12) after applying said cosmetic composition (11), or - the fade value of said cosmetic composition (11) after washing said hair (12), or - a selectivity value characterizing the color difference between the root and the tip of said hair (12); 7. The method of claim 1, comprising:

8. 8. The method according to claim 1, wherein during the collecting step (120) of the training phase (100), the set of training data (18) is collected, at least in part, via a sensor capable of measuring the at least one colorimetric parameter.

9. 9. The method of claim 1, wherein the processing phase (200) further comprises a step (230) of sending the or each colorimetric parameter and the associated test data item (19) to a display screen (16) for displaying a rendering for the or each cosmetic composition (11) under test on the display screen (16), the or each rendering representing the application of the cosmetic composition (11) under test to the hair (12).

10. 10. The method according to any one of claims 1 to 9, wherein the treatment phase (200) further comprises a step (240) of producing at least one sample of the cosmetic composition (11) under test for application on hair (12) to verify the value of the at least one colorimetric parameter.

11. 11. A computer program product having stored thereon a computer program comprising program instructions, the computer program being loaded onto a data processing unit (15) and configured to perform the method according to any one of claims 1 to 10 when the computer program is implemented on the data processing unit (15).

12. An electronic device (10) for determining at least one colorimetric parameter characterizing hair (12), in particular a cosmetic composition (11) for dyeing hair, comprising: An electronic device (10), wherein the electronic decision-making device (10) is capable of implementing the decision-making method according to any one of claims 1 to 10.

13. 1. A method for determining a target cosmetic composition (1011*) for dyeing hair (1012), in particular hair, according to at least one target colorimetric parameter (1013*), comprising: The method is performed by an electronic decision-making device (1010) and includes a processing phase (1200), the processing phase (1200) comprising: - collecting (1210) said at least one target colorimetric parameter (1013*); - obtaining (1220) a set of cosmetic compositions (1011) under test, each cosmetic composition (1011) under test comprising an ingredient (1014); - obtaining (1230) at least one colorimetric parameter (1013) associated with each cosmetic composition (1011) under test; - filtering (1235) the cosmetic compositions under test (1011) according to the at least one colorimetric parameter (1013) associated therewith to obtain filtered cosmetic compositions (1030), wherein the at least one colorimetric parameter (1013) associated with each filtered cosmetic composition (1030) satisfies a criterion for the at least one target colorimetric parameter (1013*); - forming (1240) several clusters (1035) of filtered cosmetic compositions (1030), such that the filtered cosmetic compositions (1030) from the same cluster (1035) comprise the same ingredients (1014); - selecting (1250) a cluster (1035) of the filtered cosmetic compositions (1030), called a selected cluster (1035*); - determining (1260) the target cosmetic composition (1011*) according to the filtered cosmetic compositions (1030) of the selected cluster (1035*); A method comprising:

14. each filtered cosmetic composition (1030) comprising, for each of its components (1014), an amount of said component (1014) in said filtered cosmetic composition (1030); The determining step (1260) of the processing phase (1200) - for each ingredient (1014) of the filtered cosmetic compositions (1030) of the selected cluster (1035*), calculating the average value of the amount of the ingredient (1014) from the filtered cosmetic compositions (1030); - determining, for each of said ingredients (1014), an average cosmetic composition (1040) comprising an amount equal to said calculated average value of each of said ingredients (1014); - forming the target cosmetic composition (1011*) according to the average cosmetic composition (1040); 14. The method of claim 13, comprising:

15. During the determining step (1260) of the processing phase (1200), forming the target cosmetic composition comprises: calculating at least one colorimetric parameter (1013) associated with the average cosmetic composition (1040); and optimizing the amount of each component (1014) of the average cosmetic composition (1040) such that the at least one colorimetric parameter (1013) associated with the average cosmetic composition (1040) approaches the at least one target colorimetric parameter (1013*); 15. The method of claim 13 or 14, wherein the target cosmetic composition (1011*) is the average cosmetic composition (1040) resulting from optimizing the amounts of its components (1014).

16. 16. The method according to any one of claims 13 to 15, wherein during the selection step (1250) of the processing phase (1200), M clusters (1035) of filtered cosmetic compositions (1030) are selected, M being greater than or equal to 2, and during the determination step (1260) of the processing phase (1200), for each selected cluster (1035*), a corresponding target cosmetic composition (1011*) is determined.

17. 17. The method of claim 16, wherein during the selection step (1250) of the processing phase (1200), the M clusters (1035*) are selected such that, for each selected cluster (1035*), only the ingredients (1014) contained in the filtered cosmetic composition (1030) of the selected cluster (1035*) are contained in the selected cluster (1035*).

18. The method includes a training phase (1100) prior to the processing phase (1200), the training phase (1100) comprising: - receiving (1110) a set of feasibility constraints for a cosmetic composition; - collecting (1120) a set of training data (1018), each training data item (1018) being specific to a cosmetic composition (1011) that complies with said set of feasibility constraints received; Each training data item (1018) a set of sizes representing the amounts of each component (1014) of the cosmetic composition (1011); the value of said at least one colorimetric parameter associated with said cosmetic composition (1011); and and - training (1130) an artificial intelligence model based on said set of training data (1018) to obtain a trained model; Including, 18. The method according to any one of claims 13 to 17, wherein during the step (1230) of obtaining at least one colorimetric parameter (1013) associated with each cosmetic composition under test of the processing phase (1200), each colorimetric parameter (1013) is obtained by applying the trained model to a corresponding cosmetic composition (1011) under test.

19. The step (1220) of obtaining a set of cosmetic compositions under test in the processing phase (1200) comprises: - receiving a number N corresponding to the number of cosmetic compositions (1011) under test, said number N being greater than or equal to 2; - generating N test data items (1019), each test data item (1019) relates to a cosmetic composition (1011) whose ingredients (1014) are selected from a predefined list of ingredients (1014), such that the cosmetic composition (1011) associated with each test data item (1019) is within a reduced feasible cosmetic composition space, the reduced space comprising only those cosmetic compositions (1011) that comply with a set of feasibility constraints; the N test data items (1019) represent the reduced feasible cosmetic composition space; a set of said cosmetic compositions under test (1011) is formed from said generated N test data items (1019); 19. The method of any one of claims 13 to 18, comprising:

20. During the processing phase (1200), the step of generating N test data items (1019) in the obtaining step (1220) includes a step of calculating, for each test data item (1019), a size of the N test data items (1019) such that a distance between the test data item (1019) and other test data items (1019) is maximum; 20. The method of claim 19, wherein each magnitude of each test data item (1019) represents the amount of ingredient (1014) in the corresponding cosmetic composition (1011) under test.

21. The at least one colorimetric parameter (1013) is - a triplet of values ​​(L*, a*, b*, R, G, B) characterizing the color of the hair (1012) after applying said cosmetic composition (1011), or - the fading value of said cosmetic composition (1011) after washing said hair (1012), or - a selectivity value characterizing the color difference between the root and tip of said hair (1012); 21. The method of any one of claims 13 to 20, comprising:

22. 22. The method of any one of claims 13 to 21, wherein the processing phase (1200) further comprises a step (1280) of producing at least one sample of the target cosmetic composition (1011*) for application on hair (1012) to verify the value of the at least one colorimetric parameter (1013).

23. 23. A computer program product having stored thereon a computer program comprising program instructions, said computer program product being configured to perform the method of any one of claims 13 to 22 when loaded onto a data processing unit and when implemented on said data processing unit.

24. 23. An electronic device (1010) for determining a target cosmetic composition (1011*) for dyeing hair (1012), in particular hair, according to at least one target colorimetric parameter (1013*), said electronic determination device (1010) comprising a processing unit capable of performing the method according to any one of claims 13 to 22.

25. Hair (2012), especially the initial cosmetic composition intended to dye hair (2011 i ) so that the value (Va) of at least one colorimetric parameter of the adapted cosmetic composition (2011*) corresponds to a target value (Vc), The method is performed by an electronic adaptive device (2010) and includes a processing phase (2200), the processing phase (2200) comprising: - obtaining (2210) a target value (Vc) for said at least one colorimetric parameter; - an initial cosmetic composition (2011) comprising a plurality of ingredients (2014) and an associated amount of each ingredient (2014); i (2220) obtaining data relating to the - The initial cosmetic composition (2011 i 22. A step (2230) of obtaining an initial value (Vi) of said at least one colorimetric parameter for said initial cosmetic composition (2011) on hair (2012), said initial value (Vi) being i ) resulting from a measurement after application of the initial value (Vi), which differs from the target value (Vc) by a first deviation; - adjusting the initial cosmetic composition (2011) to compensate for the first deviation between the initial value (Vi) and the target value (Vc); i ), wherein the adaptation comprises: The initial cosmetic composition (2011 i Varying the amount of each component (2014) in o evaluating the effect of each variation in the value of the at least one colorimetric parameter using an automated decision model, the automated decision model having been previously trained to determine the value of at least one colorimetric parameter according to Cosmetic Composition (2011); Including, the applied cosmetic composition (2011*) is obtained after the application; A method comprising:

26. 26. The method of claim 25, wherein the adapted cosmetic composition (2011*) has an adapted value (Va) for the at least one colorimetric parameter, the adapted value (Va) resulting from a measurement after application of the adapted cosmetic composition (2011*) on hair (2012), and the adapted cosmetic composition (2011*) is such that a second deviation between the adapted value (Va) and the target value (Vc) is less than the first deviation.

27. The processing phase (2200) - using the automated decision model to determine the initial cosmetic composition (2011) by varying the amount of each component (2014) of the initial cosmetic composition; i calculating (2240) a matrix, called the Jacobian matrix (J), that represents the variation of the values ​​of the at least one colorimetric parameter of further comprising 27. The method of claim 25 or 26, wherein during the evaluation of the impact of each variation in the value of the at least one colorimetric parameter in the adaptation step (2250) of the processing phase (2200), the impact is evaluated by calculating the Jacobian matrix (J) and a value of a cost function (FC) that depends on each variation.

28. The cost function (FC) conforms to the following formula: FC=||Vi-Vc+JΔmp|| where FC is the cost function, Vi is the initial value of the at least one colorimetric parameter; Vc is the target value; J is the Jacobian matrix, 28. The method of claim 27, wherein Δmp is a vector comprising the amount of variation of each component (2014).

29. 29. The method according to any one of claims 25 to 28, wherein during the adaptation step (2250) of the processing phase (2200), the variation and the evaluation are repeated several times by an optimization algorithm.

30. During the adaptation step (2250) of the processing phase (2200), the optimization algorithm applied is adapted to a set of feasibility constraints (E cv ) is an optimization algorithm constrained by The initial cosmetic composition (2011 i ) and the adapted cosmetic composition (2011*) is determined by the set of feasibility constraints (E cv 30. The method of claim 29, wherein

31. The method includes a training phase (2100) prior to the processing phase (2200), the training phase (2100) comprising: - receiving (2110) a set of feasibility constraints for the cosmetic composition; - collecting (2120) a set of training data (2018), each training data item (2018) being specific to a cosmetic composition (2011) that complies with the set of feasibility constraints received; Each training data item (2018) is A set of sizes representing the amounts of each component (2014) of the cosmetic composition (2011); the value of said at least one colorimetric parameter associated with said cosmetic composition (2011); and - training (2130) an artificial intelligence model based on the set of training data (2018) to obtain a trained model capable of determining, from a cosmetic composition, a value of the at least one colorimetric parameter of the cosmetic composition; Including, 31. The method of any one of claims 25 to 30, wherein during the adapting step (2250) of the processing phase (2200), the model for determining at least one cosmetic composition colorimetric parameter is the model trained during the training step (2130) of the training phase (2100).

32. 32. A method according to any one of claims 25 to 31, wherein the processing phase (2200) further comprises a step (2270) of producing a sample of the adapted cosmetic composition (2011*) to measure the adapted value (Va) of the at least one colorimetric parameter of the adapted cosmetic composition (2011*).

33. The processing phase (2200) is repeated at least once to form at least two iterations; During the step (2220, 2230) of obtaining the or each iteration of the processing phase (2200), the initial cosmetic composition (2011 i 33. The method according to any one of claims 25 to 32, wherein the initial value (Vi) results from a measurement after the application of the adapted cosmetic composition (2011*) from the previous iteration on hair (2012).

34. 34. A computer program product having stored thereon a computer program comprising program instructions, said computer program product being configured to perform the method of any one of claims 25 to 33 when loaded onto a data processing unit and when implemented on said data processing unit.

35. 1. An electronic device (2010) for adapting an initial cosmetic composition intended for dyeing hair, in particular hair, so that the value of at least one colorimetric parameter of said adapted cosmetic composition corresponds to a target value, said device comprising: An electronic device (2010), wherein said electronic adaptation device (2010) is capable of implementing the adaptation method according to any one of claims 25 to 33.

Images