A method for obtaining an optical lens with a target hue.

Abstract

The present invention relates to a method for obtaining an optical lens having a target hue, comprising the steps of: - providing a process for coloring an optical lens to a target hue and a set of process variables suitable for changing the target hue, each process variable being able to take a value; - providing several optical lens samples that are colored according to the above process for coloring and associated with the set of values ​​of the process variables; - measuring the measured hue of each optical lens sample; - providing a model that takes a set of values ​​for the process variables as input and gives a modeled hue for an optical lens colored by the above process as output; - fine-tuning the model by minimizing a first function that depends on the difference between the measured hue of an optical lens sample and a second modeled hue; - inverting the model; and - executing the process using the set of values ​​to obtain an optical lens having a target hue.

Description

[Technical Field]

[0001] This invention generally relates to the field of eyeglasses.

[0002] The present invention relates, more particularly, to the coloring of lenses. [Background technology]

[0003] It is well known that tinting the lenses of sunglasses and eyeglasses is done to protect the wearer's eyes from the sun's radiation.

[0004] For this purpose, most optical lens manufacturers provide lens catalogs with a limited number of colors. Whenever a new color is proposed in the catalog, several tests (up to eight) are required to obtain the desired shade.

[0005] To increase the freedom of choice regarding lens tint, wearers can be offered the option of showing their preferred color by bringing a sample of that color, or by selecting from a very large palette.

[0006] The main drawback in this case is that it is extremely difficult for the manufacturer to accurately reproduce the desired color. In this situation, the manufacturer often has to remanufacture the lens several times until the desired hue is achieved, which is costly and time-consuming.

[0007] Japanese Patent Publication No. 2018-191073 also provides a neural network-based method suitable for learning what the best coloring recipe is for coloring lenses to satisfy the wearer, based on a set of numerous training examples.

[0008] The drawback of this method is that it requires a very large amount of data to fine-tune the neural network, which is time-consuming and expensive. Furthermore, such neural networks form a black box that is difficult to verify and can sometimes produce very erroneous results. [Overview of the project] [Problems that the invention aims to solve]

[0009] In this regard, the present invention provides a simple and accurate solution for reproducing a color selected by the wearer in a lens. [Means for solving the problem]

[0010] More precisely, the method according to the present invention (for obtaining an optical lens having a target hue) is: - A step of providing a process for coloring an optical lens to a target hue and a set of process variables suitable for changing the target hue, wherein each process variable can take a value selected from a group of possible values. A1) A step of providing several optical lens samples that have been colored according to the above process for coloring, wherein each optical lens sample is related to a set of values ​​(those used in the process to obtain the optical lens samples), A2) A step of measuring the measured color of each optical lens sample, - A step of providing a first model that takes a set of values ​​for process variables as input and gives a first modeled hue for an optical lens colored by the process as output, - A step of determining a first modeled hue for each optical lens sample by using the first model with the set of values, - A step of providing a second model that takes a set of values ​​for process variables as input and gives a second modeled hue for an optical lens colored by the process as output, A3) A step of fine-tuning the second model by minimizing a first function that depends on the difference between the measured hue and the first modeled hue for an optical lens sample, - A step of solving an inverse problem to invert the second model described above, wherein the inverted second model takes a target hue as input and gives as output a usable set of values ​​for process variables that produce the target hue; and a step of performing a process using the usable set of values ​​to obtain an optical lens having the target hue. Includes.

[0011] In other words, the present invention proposes using a model to determine the best recipe for obtaining a lens with a target hue.

[0012] Here, the second model is estimated from the first model, meaning it differs by a difference referred to below as the "error function." As a result, step A3 makes it possible to determine this error function in order to find the best second model. Thus, the solution according to the present invention makes it possible to fine-tune this second model to always find the best recipe based on the samples. Therefore, it is possible to find an accurate model with fewer samples.

[0013] Another advantage of this solution is that, due to its simplicity, it is possible to understand how this model works and to easily determine how accurate it is under any circumstances.

[0014] As a result, it becomes easier to suggest new colors in the catalog and to provide new shades on demand.

[0015] Other preferred features of the present invention are as follows: - The first model is based on a linear combination of process variable values, the values ​​of which are preferably related to dye concentration. - The first model described above is based on the Lambert-Beer law. - The hue is characterized by the hue parameter. - The second model is such that each tint parameter characterizing the second modeled tint depends on the sum of the corresponding tint parameter characterizing the first modeled tint and an error function, and the first function also depends on the error function. - Each error function includes the sum of terms, and each term includes the multiplication coefficient determined in step A3). - The error function contains a predetermined polynomial for each of the hue parameters, and the terms of the predetermined polynomial contain process variables as variables. - In order to establish the error function, a regularization method is applied to the predetermined polynomial to identify the most relevant term and at least one term to be removed from among the terms of the predetermined polynomial, and the error function is then determined using only the relevant term. - The above calculation is performed for each of the measured shades. - The error function includes a predetermined polynomial that uses parameters that characterize the process variables as variables. - The error function includes the sum of a quadratic polynomial and a term that depends on the product of at least two parameters that characterize the process variables. - The optical lens samples are divided into at least two separate groups that do not share any common elements. - Step A3) is performed on the first of the group to determine the error function. - The method includes the steps of: A41) obtaining a set of values ​​for each optical lens sample belonging to the second of the group; A42) measuring the measured hue of each optical lens sample belonging to the second of the group; A43) predicting the hue of each optical lens sample belonging to the second of the group based on the second model and the set of values ​​obtained in substep A41); and A44) verifying the error function according to the difference between the measured hue and the predicted hue. - In step A1), all sets of values ​​are different from each other. - The step is repeated for several groups of several samples, and for each group, a new error function is established as a function of a previously determined error function, which is distinct and preferably has no common elements with each other, in order to determine several error functions.

[0016] In a preferred embodiment, the method is E0) The user selects a color sample, E1) An operation to determine the target hue of the color sample, E4) An operation to obtain a second model in accordance with any one of the preceding claims, and to calculate a corresponding set of values ​​for process variables based on the second model and the target shades using inverse problem optimization of the second model. Includes.

[0017] Other preferred features of the present invention are as follows: - After operation E4), a color sample is formed by an optical lens or lens template, and steps A1) to A3) are repeated, taking into account the optical lens sample and the color sample. - During operation E1), the target color is determined by the image sensor within the visible light range. - During operation E1), the target hue is determined based on several images of the aforementioned sample of color viewed from different angles and / or under different lighting conditions. - During operation E1), the target hue is determined by acquiring images of the color sample and the color model, and by processing the images so that the hue of the color sample appearing on the processed image corresponds to the hue of the color model.

[0018] The present invention also relates to a method for manufacturing optical lenses, - Identification of the coloring recipe using the method described above, - Manufacturing of the colored optical lens by coloring the colored optical lens using the set of values ​​calculated in operation E4) and This relates to a method for manufacturing optical lenses, including [specific details omitted].

[0019] The present invention and how it may be carried out will become clear from the following description with reference to the attached drawings, which are provided as non-limiting examples. [Brief explanation of the drawing]

[0020] [Figure 1] This is a diagram showing the process according to the present invention. [Modes for carrying out the invention]

[0021] The present invention aims to characterize a color and reproduce that color in an optical lens.

[0022] In the following, the terms "color" and "hue" are used to refer to the same concept, with the term "hue" being used more specifically to describe the color tone of a nearly transparent lens. Furthermore, in the following explanation, we assume that color can be described by hue parameters (e.g., hue, saturation, and lightness).

[0023] In order to carry out the present invention, a device (such as a camera) is required to acquire an image of a color selected by an individual (hereinafter referred to as the wearer, in the sense that the lens is attached to the eyeglass frame worn by the individual).

[0024] For simplicity's sake, let's assume this camera is located in an eyeglasses store.

[0025] A control unit is also needed to process these images. This control unit comprises a processing unit such as a CPU, memory, and various input / output interfaces.

[0026] The processing unit is well-suited for receiving acquired images because it has its own input interface.

[0027] The processing unit, equipped with its own output interface, is suitable for transmitting new images to display screens viewed by opticians and wearers. The processing unit is also programmed to transmit data to colored lens manufacturing devices.

[0028] Since the processing unit has its own memory, it stores computer applications, which are computer programs containing instructions, and through the execution of instructions by the processor, the processing unit can carry out the methods described below.

[0029] First, this method involves finding a model (hereinafter referred to as the second model M2) that matches the lens manufacturing recipe with the lens's color tone.

[0030] This second model M2 is associated with a specific process P0 for coloring optical lenses. In other words, this model is a reliable recipe only if the lenses are manufactured according to a specific process P0. Consequently, the manufacturing process P0 includes substantially immutable steps that must be reproduced. For this reason, the preliminary action is to define this process.

[0031] Naturally, within these process steps, parameters ("process variables" or process parameters p) are used to change the tint of the resulting lens. i It is possible to change the value of (also known as)

[0032] Here, process P0 is based on sublimation technology.

[0033] This process includes a first step of depositing ink onto transfer paper according to a predetermined pattern by inkjet printing, wherein each dye is deposited and spread according to a planned hue and / or hue pattern.

[0034] Next, in the second step, the surface of the lens to be colored is placed facing the transfer paper at a distance from it, and the transfer paper is heated to sublimate the dye, thereby transferring the dye to the surface of the lens.

[0035] Finally, in the third step, the dye is fixed onto the lens, for example, by penetration into the material.

[0036] Several exemplary embodiments are disclosed in European Patent No. 1637313, U.S. Patent No. 6534443, or U.S. Patent No. 6840621.

[0037] In a modified form, process P0 may be based on other techniques. For example, the modified form may involve directly inking the surface of the lens, coloring the lens in chunks (as a whole material), or coloring the lens by surface impregnation.

[0038] Parameter p characterizing process P0 i This may include parameters related to the hue itself or the process itself, such as the type of transfer paper and the distance between the transfer paper and the lens.

[0039] In the embodiment used as an application example, the parameters are considered to consist only of the proportions of components used to color the lens. In this embodiment, the number of parameters is equal to 3, meaning that only three dyes are used to color the lens. One of these parameters is the amount R of magenta dye in the material used to form the lens. Another parameter is the amount Y of yellow dye, and the last parameter is the amount B of cyan dye. Naturally, in a variant, another triplet of colors may be used to color the lens. In another variant, the number of dyes may be greater (for example, some black, or two shades of red, or two different cyan / blue shades may be used).

[0040] Each ratio R, Y, B, i.e., each process parameter p i The value is selected from a group of possible values, where each quantity corresponds to the dose used by the machine. Thus, each process parameter p i v is a value expressed as an integer, or as a percentage in dpi, or concentration, or any other unit related to process P0. i It may have.

[0041] Process P0 is defined by each process parameter p i value v i Once determined, it is parameterized as described above.

[0042] Before describing process P0 in more detail, we can first explain the second model M2 and how it is configured.

[0043] Here, the solution is the value v i This is based on a first model that can match a set of colors to a modeled shade.

[0044] In this regard, please note that several solutions can be used to demonstrate the characteristics of the color tones.

[0045] In the embodiment under consideration, the reason why the CIELAB color space is selected instead of other color spaces is not only because the color distribution is simple, uniform, and homogeneous, but also because of the recognized effectiveness of CIELAB color difference.

[0046] In this color system, a color can be defined by three data, namely, the lightness or brightness L (which is 0 in the case of black and 100 in the case of white), and two coordinates a and b. These two coordinates a and b are related to the hue and chroma (or saturation) of the color.

[0047] Here, the first model M1 is more specifically established so that the values of the color parameters L, a, b can be calculated according to the value v i (R, Y, B) of the process parameter p i .

[0048] This first model M1 matches these values in a simple and sometimes inaccurate way.

[0049] In this embodiment, this first model M1 is established according to the Lambert - Beer law.

[0050] According to this first model, the color parameters L, a, b form a predetermined function L i , a M1 , b M1 that changes according to the set of values v M1 .

[0051] These functions are known, but can also be set by using a preliminary sample in which only one value in the set of values v i (R, Y, B) of the process parameter p i is not null.

[0052] More precisely, at least three preliminary lens samples made solely of magenta dye with three different intensity values ​​(where the intensity value can be concentration or the number of droplets per pixel) can be used, at least three preliminary lens samples made solely of yellow dye with three different intensity values, and at least three preliminary lens samples made solely of cyan dye with three different intensity values.

[0053] Each triplet of samples allows us to determine a linear equation that expresses the variation of each hue parameter L, a, and b as a function of the variation in the concentration of each dye, and a threshold that, when exceeded, causes the hue parameter to saturate regardless of the concentration value above that threshold.

[0054] The inventors found that while the Lambert-Beer law works well for dealing with the hues of liquids, it is not very accurate for describing the complex coloring process of lenses. In fact, the process may have different nonlinear migration kinetics for each dye, and / or interactions between dyes may change the migration kinetics of one or more dyes. Therefore, it is important to define a second model M2 in order to find a more accurate recipe. This second model M2 is based on the first model, where the hue parameters L,a,b are replaced with the process parameter p i Setting value v i It has been improved to match more accurately.

[0055] This second model uses the function L emitted from the first model M1. M1 ,a M1 ,b M1 and error function Error L Error a Error b Accordingly, it is defined for each function of model M1.

[0056] More precisely, according to this second model M2, each hue parameter L M2 ,a M2 ,b M2This is the corresponding hue parameter L emitted from the first model. M1 ,a M1 , or b M1 And the corresponding error function Error L Error a Error b It is equal to the sum of the two.

[0057] It can be written as follows: L M2 (v i )=L M1 (v i )+Error L (v i ) a M2 (v i )=a M1 (v i )+Error a (v i ) b M2 (v i )=b M1 (v i )+Error b (v i )

[0058] Each error function is related to the process parameter p i The value of (R,Y,B) v i Please note that this depends on the set of elements.

[0059] In this embodiment, each error function is a polynomial. More precisely, each error function is equal to the sum of terms, where each term is at least one value v i The multiplication coefficients applied are α0, α1, ..., α 10 Includes.

[0060] In one embodiment, the polynomial error function is based on a quadratic polynomial function, and they may further include one or more cubic or quartic components, particularly components that take two, three, or more process parameters together.

[0061] It can be written as follows: Error L(v i ) = α0 L + α1 L ·R + α2 L ·Y + α3 L ·B + α4 L ·R 2 + α5 L ·Y 2 + α6 L ·B 2 + α7 L ·R·Y + α8 L ·R·B + α9 L ·Y·B + α 10 L ·R·Y·B Error a (v i ) = α0 a + α1 a ·R + α2 a ·Y + α3 a ·B + α4 a ·R 2 + α5 a ·Y 2 + α6 a ·B 2 + α7 a ·R·Y + α8 a ·R·B + α9 a ·Y·B + α 10 <000008⑨>·R·Y·B Error b (v i ) = α0<00000⑨2>+ α1<00000⑨3>·R + α②<00000⑨4>·Y + α3<00000⑨5>·B + α4<00000⑨6>·R[[ID=8③]]<00000⑨7>+ α5<00000⑨8>·Y<00000⑨9>+ α6<00001⑩0>·B<00001⑩1>+ α7 <00001⑩2>·R·Y + α8<00001⑩3>·R·B + α9<00001⑩4>·Y·B + α<00001⑩5><00001⑩6]·R·Y·B

[0062] In these equations, note that R, Y, and B form a set of values v of the process parameter p i of i It should be noted that a set is formed. It should be noted that there are some special symbols in the original text that may need to be further clarified in the context. Also, the numbering and formatting in the translation try to maintain consistency with the original as much as possible. If there are any specific requirements or corrections regarding the translation, please feel free to let me know.

[0063] The objective is to find the best equation for the three error functions by multiplying coefficient α0 L ~α 10 L ,α0 a ~α 10 a ,α0 b ~α 10 b The task is to fine-tune these. In fact, once these coefficients are determined, the second model M2 is clearly defined, and these values ​​v i Whatever the value v i Set of color parameters p i It is considered possible to make it match.

[0064] The following process P1 is used to fine-tune the multiplication coefficient value (see Figure 1).

[0065] This process P1 involves the optical lens sample S. j This includes a preliminary step A0 to obtain a set of lens samples. This set preferably includes more than 50 lens samples. Here, this set includes 220 lens samples S. j Includes.

[0066] These lens samples S j This is two groups G with no common elements. i and G i’ The samples are divided into groups. These groups may contain a distinct number of samples, but here, each group contains exactly half the total number of samples before the division.

[0067] Lens sample S j Group G (first group) i This is the multiplication coefficient α0 of the error function, which will be explained in more detail below. L ~α 10 L ,α0 a ~α 10 a ,α0 b ~α 10 bUsed to determine the other group G. i’ This is used to check whether the obtained error function is accurate.

[0068] To fine-tune these multiplication coefficients, the process is performed on each lens sample S of the first group. j Regarding the process parameter p i value v i,j This includes a first step A1 to obtain (where the subscript j corresponds to the lens sample being considered). For example, since the lens sample is manufactured according to the selected manufacturing process P0, these values ​​can be considered known.

[0069] Next, the process is for the first group G i Each lens sample S j Regarding the value of the hue parameter L j ,a j ,b j This includes a second step A2 in which these values ​​are obtained, where they are measured more precisely. This step is performed, for example, using a spectrophotometer in the visible range of light.

[0070] The third step A3 is the multiplication coefficient α0 L ~α 10 L ,α0 a ~α 10 a ,α0 b ~α 10 b The measured value L j ,a j ,b j and Group G i Lens sample S j Process parameter p i value v i,j This involves evaluating it as a function of .

[0071] For this purpose, the first function f L ,f a ,f b It will be introduced.

[0072] These first functions f L ,f a ,f b This is the measured value L j ,a j ,b j And the value L issued from the first model M1 M1 ,a M1 ,b M1 (Each lens sample S) j It depends on the difference between (what must be calculated for) and

[0073] More precisely, here, these first functions f L ,f a ,f b This is the value L measured for each lens sample. j ,a j ,b j and the value L issued from the second model M2 M2 ,a M2 ,b M2 It depends on the difference between [the two values].

[0074] It can be written as follows:

number

[0075] Next, the first group G of the sample i The best multiplication coefficient α0 L ~α 10 L ,α0 a ~α 10 a ,α0 b ~α 10 b To find the following function f L ,f a ,f b We minimize each of these. This is a linear least squares problem, and therefore the linear system is solved.

[0076] Next, process P1 includes a fourth step A4 to verify the obtained error function. This step A4 includes four substeps.

[0077] The first substep a41) is the second group G i’ Lens sample S belonging to this category j Process parameter p i The value of (R,Y,B) v i,j The goal is to obtain a set of values.

[0078] The second substep a42) involves these lens samples S j The values ​​of the color parameters L, a, and b. j ,a j ,b j The purpose is to measure.

[0079] The third substep a43) is the second model M2 (first group G i Based on the sample settings, and the process parameter p obtained in substep aA41) i value v i,j Using these lens samples S j The values ​​of the color parameters L, a, and b. M2 ,a M2 ,b M2 It is about prediction.

[0080] The fourth substep a44) is the value L predicted in substep a43). M2 ,a M2 ,b M2 and the value L measured in substep a42) j ,a j ,b j Accordingly, the objective is to evaluate the accuracy of the aforementioned second model M2.

[0081] In this fourth substep a44), the quality assessment of the second model M2 may be performed by an ECMC metric test. This test can assess whether the value predicted in substep a43) and the value measured in substep a42) are sufficiently close.

[0082] According to this test, the difference ΔE between these values ​​can be calculated using the following function.

number

[0083] In the above equation,

number

number

[0084] In other words, this test is for the second group G i’ For each lens sample, the difference ΔE is used to determine the first group G i The accuracy of the second model established using the lens samples is measured.

[0085] If all of the calculated differences fall below a predetermined threshold, then the second mode discovered is considered to be the determined error function (i.e., the multiplication coefficient α0). L ~α 10 L ,α0 a ~α 10 a ,α0 b ~α 10 b ) is accurate enough to maintain the set second model M2.

[0086] While the calculation of ΔE has been described here, it should be noted that this disclosure may also be applied to determine the coefficients of the error function by solving a least-squares problem of a function F that depends on all parameters and coefficients of the error function and aims to minimize the sum of ΔE and the first model data. Furthermore, other minimization functions may be used to determine the coefficients of the error function.

[0087] In this case, we solve a nonlinear least-squares problem corresponding to minimizing the following function F.

number

[0088] In one embodiment, the preceding step is to determine several error functions and associated precision (difference ΔE) using a sample S j Several groups of several samples selected within the same set G i This is repeated using or using other sets of samples.

[0089] In this embodiment, group G of the sample i , G i’ It is selected from the 220 samples before splitting.

[0090] Each group G i , G i’ The selection of samples in Group G i , G i’ This is different from the one previously selected in another set of groups G. i , G i’ This includes 110 samples, which are, for example, selected randomly.

[0091] The step is repeated at least 10 times, and here it's repeated 50 times.

[0092] Finally, the process unit defined 50 error function triplets and 50 exact ratios.

[0093] Using these 50 triplets, it is possible to determine whether some terms in each error function can be removed to simplify and / or make the calculations more accurate.

[0094] For this purpose, fine-tuning methods are applied to these error functions.

[0095] For example, a second model approximation can be constructed based on different groups of lens samples and a triplet of the error function, and the final model can be constructed by comparing the different approximations and removing the coefficients that differ most significantly between them. These coefficients are actually considered to be excessively "noise-related." The coefficients to be removed are, for example, the N most different multiplication coefficients (where N is a given integer), or those whose difference is greater than a predetermined threshold.

[0096] In this step, the second model M2 is group G of the samples. i It is recalculated based on, but the error function has some coefficients already set to a value of 0, and minimization is performed only for the other coefficients. Therefore, without taking into account the variable coefficients, the process parameter p i The value of (R,Y,B) v i Based on this set, it is possible to calculate the hue parameters L, a, and b.

[0097] Alternatively, the fine-tuning step is Sample G i This can be replaced by a regularization method that uses only one group.

[0098] Therefore, with or without fine-tuning, this disclosure makes it possible to define a second model that enables the prediction of hue parameters based on input process parameters.

[0099] It should be noted that this second model is preferably considered valid for only one lens substrate material. In other words, the coefficients of the second model must be calculated for each material. Conversely, the first model is preferably independent of the substrate material (however, in a modified form, the first model may also depend on the substrate material).

[0100] Conveniently, this second model can be inverted, resulting in a process parameter value v based on the target hue parameters L, a, b. i It is possible to determine the set of parameters. This can be done using an optimization process, such as Newtonian methods.

[0101] Next, it is possible to determine the color to be applied to the lens and then carry out the following process to manufacture a lens having this hue.

[0102] This process involves several steps that are executed sequentially.

[0103] Preliminary step E0 is for the wearer to select the color.

[0104] This step can be carried out in various ways.

[0105] For example, the wearer may bring an element (such as a piece of cloth or paper) that has the selected color.

[0106] Alternatively, the wearer can go to their own optician and choose eyeglass frames. In this case, the control unit allows input of criteria for the frame (or its color) so that it can provide a limited selection of available shades. As a result, the shade is selected from a limited number of colors (including the frame color) that match the frame. The wearer can then request a change to this color, for example, by using appropriate software.

[0107] An optician may also use pre-tinted lenses to determine the chosen shade.

[0108] At this stage, the color selected by the wearer is likely not yet manufactured by the optical lens manufacturer.

[0109] Otherwise, as will be described later (Step E6), the recipe for coloring the lenses is already known and can therefore be read and applied directly.

[0110] Next, the first step E1 is to describe the characteristics of this color, which we will call the target hue.

[0111] In other words, the goal is the values ​​of the target color parameters L, a, and b L. t ,a t ,b t The task is to make a decision.

[0112] If the wearer selects the color via software, this decision is easy to implement.

[0113] However, if the wearer brings an element with the desired color, this decision will be made in a different way.

[0114] For this purpose, the color of the brought-in element is acquired via a camera system. For this purpose, the control unit displays an image of the sample and a target indicating the position where the color is determined on the screen.

[0115] The size and position of this target may be modified by the optician, for example, to adjust its size to match the overall size of the elements brought in.

[0116] In a preferred embodiment, several images of the brought element are taken from different angles and / or under different lighting conditions, and the target color of the brought element is calculated based on all these images (e.g., by averaging or by careful or corrected calculation).

[0117] In the modified configuration, or in addition, a reference element with a known color is placed next to the brought-in element so that both elements are visible to the camera system. The reference element is, for example, a colored piece of paper with a known Pantone code. The control unit can then determine the color of the reference element on the acquired image and correct the overall image color so that the color of the reference element exactly matches the known color. Finally, the color of the brought-in element can be determined on this corrected image.

[0118] In the transformed form, the reference element can be observed near a distinct color that exists in a distinct region, such as a known and known intensity, for example, a primary color. Therefore, the exact color of the brought element can be determined more accurately.

[0119] It should be noted that the elements brought in may have various properties and structures, and may be transparent or opaque. As a result, their appearance may be matte, satin, or glossy, or even transparent or opaque. Therefore, in modified forms, or in addition, the process unit can determine the light reflectance of the brought in element and correct the image obtained as a function of reflectance.

[0120] In the transformed form, or in addition, by using a separate light-emitting element, it is possible to determine the exact color of the brought-in element.

[0121] In fact, the type of environmental emitters can be considered, for example, by a declarative system that takes into account the average reference spectrum or any useful data, as well as the material of the elements. As a result, the emitters used during color acquisition are registered, and the color interpretation is performed by subtracting the emitter spectrum from the measured value. In a modified form, to overcome the effects of metamerism, several reference emitters are applied to the acquired image and standard to measure their influence and sensitivity to color.

[0122] In the transformed form, or in addition, it is possible to automatically adjust the acquired color, especially if the brought-in element is solid, to convert it to a semi-visible color (i.e., a color that colors a transparent lens).

[0123] The second step, E2, is for the wearer, or an optician or another ophthalmic specialist, to change the acquired color to take into account the wearer's expected results (and any technical limitations).

[0124] For this purpose, the wearer or optician may use appropriate software to display the acquired color on a screen, allowing the wearer to change the target hue (i.e., at least one of the hue parameters, i.e., hue, saturation, or lightness).

[0125] Once the target hue is determined, in the third step E3, the control unit sets the value L of the hue parameter for the target color. t ,a t ,b t Calculate.

[0126] As described above, in the embodiments described, these parameters are expressed in the CELIAB system, but in variations, they may be expressed in a different system (for example, the RGB system).

[0127] Next, the control unit will use these values ​​L t ,a t ,b t It checks whether the entry is already known within that database.

[0128] In the following, these values ​​L t ,a t ,b t We consider this to be something that is not known.

[0129] It should be noted that the control unit may reside at the optician's location, or on a remote server, such as a server at the lens manufacturer's location or a server on the cloud.

[0130] Next, in the fourth step E4, the control unit takes these values ​​and issues the instructions necessary to reproduce the color on the optical lens, namely the process parameter p i value v i,t Calculate the set (R, Y, B).

[0131] For this purpose, the control unit uses a second model M2, or more precisely, an optimization method to invert the second model. Note that this control unit is not necessarily the same as the one used previously. This unit preferably belongs to the lens manufacturer. Therefore, it may be a different control unit from the one used in steps E2 and E3, or it may be the same control unit if step E3 or even E2 is performed on the manufacturer's side.

[0132] Next, in the fifth step E5, the optical lens is set to the value of the process parameter v i,t It is manufactured using process P0 selected from the above set.

[0133] As explained above, in this step, the coloring of the lens can be carried out in various ways, for example, by inkjet, sublimation transfer, or absorption.

[0134] After the lens is manufactured, in the sixth step E6, a value L is set that represents the target color characteristics. t ,a t ,b t and value v i,t The corresponding sets can be recorded in the database.

[0135] To improve the second model, it is actually possible to measure the tint of the resulting lens (since the second model is not necessarily perfect, it may be very close to the target color but slightly different), and then use this resulting lens as a new lens sample (the pool of lens samples is thus gradually strengthened). In other words, the pool of lens samples can be strengthened while generating new tints, and the second model is periodically recalculated based on this strengthened sample pool.

[0136] The present invention is by no means limited to the embodiments described and illustrated.

[0137] In particular, the process involves three process parameters p, where R, G, and B are amounts related only to the dyes used in the process. i It is characterized by the following. However, in variants, other process parameters, such as local water content measurements, temperature, pressure, type of sublimation support paper, and reference number of the machine used, may be considered. In these variants, if the location or machine used to carry out the manufacturing process P0 is changed, the entire method may be reapplied to reset the second model M2. If the parameters of process P0 that were not considered in the initial formation of the second model M2 are changed, the entire method may be reapplied to reset the second model M2, but this can be done using an equation that integrates further parameters other than R, G, and B. In variants, if these process parameters that were not initially considered are taken into account, the principles of this disclosure may be used to determine the new model M3 using the previously determined second model M2 instead of the reference model M1.

Claims

1. Target color (L ｔ , a ｔ , b ｔ A method for obtaining an optical lens having the following characteristics: Process for coloring optical lenses (P ０ ) and a process variable (p) suitable for changing the color tone of the optical lens. ｉ A step of providing a set of process variables (p ｉ ) is a value selected from the group of possible values ​​(v ｉ ) can be taken, steps, A1 - the process (P ０ ) of providing several optical lens samples (S ｊ ) colored according to, each optical lens sample (S ｊ ) being related to a set of values (v ０ ) of the process variables (p ｊ ) used to obtain the optical lens sample (S ｉ ) in the process (P ｉ，ｊ ), the step and, A2 - Samples of each optical lens (S ｊ ) Measured color (L ｊ , a ｊ , b ｊ The steps include measuring ) and The aforementioned process variable (p ｉ The process (P) receives a set of values ​​for ) as input and ０ A first model (M) gives as output a first modeled hue for an optical lens colored by ). １ The steps include providing, Value (v) ｉ，ｊ By using the first model in the aforementioned set of ), each optical lens sample (S ｊ The first modeled shade (L) of ) Ｍ１ , a Ｍ１ , b Ｍ１ ) and The aforementioned process variable (p ｉ A second model (M) takes a set of values ​​for ) as input and outputs a second modeled hue for the optical lens colored by the process described above. ２ The steps include providing, A3 - The optical lens sample (S ｊ The measured color (L) for ) ｊ , a ｊ , b ｊ ) and the first modeled hue (L Ｍ１ , a Ｍ１ , b Ｍ１ The first function (f) depends on the difference between ) Ｌ , f ａ , f ｂ By minimizing the above second model (M ２ ) and the step of fine-tuning, The second model (M ２ A step of solving the inverse problem to reverse the second model, wherein the reversed second model has the target hue (L ｔ , a ｔ , b ｔ ) is taken as input, and the target color (L ｔ , a ｔ , b ｔ The value (v) for the process variable that brings about the result ｉ，ｔ The steps involve solving the inverse problem, which gives the available set of ) as output, and the target color (L ｔ , a ｔ , b ｔ To obtain the optical lens having the value (v ｉ，ｔ Using the aforementioned available set of the process (P ０ The steps to perform ) Methods that include...

2. The first model (M １ ) is the process variable (p ｉ The aforementioned value (v ｉ The method according to claim 1, wherein the value is preferably related to the concentration of the dye, based on a linear combination of ).

3. The first model (M １ The method according to claim 2, based on Lambert-Beer's Law.

4. The hue is characterized by the hue parameters (L, a, b), and the second model (M ２ ) is such that each color parameter (L, a, b) characterizing the second modeled color is equal to the corresponding color parameter (L, a, b) characterizing the first modeled color and the error function (Error Ｌ Error ａ Error ｂ It depends on the sum with the first function (f Ｌ , f ａ , f ｂ ) also the aforementioned error function (Error Ｌ Error ａ Error ｂ The method according to any one of claims 1 to 3, which depends on ).

5. Each error function (Error Ｌ Error ａ Error ｂ The method according to claim 4, wherein the sum of the terms includes the multiplication coefficient determined in step A3).

6. The aforementioned error function (Error Ｌ Error ａ Error ｂ ) includes a predetermined polynomial for each of the aforementioned color parameters (L, a, b), and the terms of the predetermined polynomial are the process variables (p ｉ The method according to claim 4 or 5, comprising ) as a variable.

7. The aforementioned error function (Error Ｌ Error ａ Error ｂ In order to establish the error function (Error), a regularization method is applied to the predetermined polynomial, and the most relevant term and at least one term to be removed are identified from among the terms of the predetermined polynomial, and the error function (Error) Ｌ Error ａ Error ｂ The method according to claim 6, wherein the terms are then determined using only the terms that are most relevant.

8. The optical lens sample (S ｊ ) is divided into at least two separate groups that do not have any common elements with each other, and step A3) the error function (Error Ｌ Error ａ Error ｂ To determine, the method is performed on the first of the group, a41) For each optical lens sample belonging to the second of the above group, the value (v ｉ，ｊ The steps include obtaining the aforementioned set of, a42) Each optical lens sample (S) belonging to the second of the group ｊ The measured color (L) of the above-mentioned color ｊ , a ｊ , b ｊ The steps include measuring ) and a43) Based on the second model and the values (v ｉ，ｊ ) obtained in sub-step a41), for each optical lens sample (S ｊ ) belonging to the second one of the group, predicting the color tone (L, a, b); a44) The measured color (L ｊ , a ｊ , b ｊ The steps include verifying the error function according to the difference between ) and the predicted color (L, a, b) and The method according to any one of claims 4 to 7, including the method described in any one of claims 4 to 7.

9. In step A1), the value (v ｉ，ｊ The method according to any one of claims 1 to 8, wherein all of the aforementioned sets of ) are different from each other.

10. E0) The user selects a color sample, E1) An operation of determining the target color tone (L ｔ , a ｔ , b ｔ ) of the color sample, and E4) The second model (M ２ ) is obtained, and the second model and the target color (L) are obtained using the inverse problem optimization of the second model. ｔ , a ｔ , b ｔ Based on the above process variable (p ｉ The value (v) ｉ，ｔ The operation of calculating the corresponding set of ) The method according to any one of claims 1 to 9, including the method described in any one of claims 1 to 9.

11. The identification method according to claim 10, wherein the color sample is formed by an optical lens or lens template after operation E4), and steps A1) to A3) are repeated by considering the optical lens sample and the color sample.

12. During operation E1), the target color (L ｔ , a ｔ , b ｔ The identification method according to claim 10 or 11, wherein the is determined by an image sensor in the visible range of light.

13. During operation E1), the target color (L ｔ , a ｔ , b ｔ The identification method according to any one of claims 10 to 12, wherein the color is determined based on several images of the sample of color viewed from different angles and / or under different lighting conditions.

14. During operation E1), the target color (L ｔ , a ｔ , b ｔ The identification method according to any one of claims 10 to 13, wherein the color is determined by acquiring images of the color sample and the color model, and by processing the images such that the hues (L, a, b) of the color sample appearing on the processed images correspond to the hues of the color model.

Images