Techniques for color matching

By combining a spectrophotometer and a physical information machine learning algorithm with the radiation transfer function, a set of color components is generated, which solves the problems of low efficiency and inaccuracy in existing color matching technologies and achieves efficient and accurate coating color matching.

CN118922696APending Publication Date: 2024-11-08PPG INDUSTRIES OHIO INC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202380029307.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-12-19
Filing Date
2023-03-24
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

Existing color matching technologies are inefficient, time-consuming, and inaccurate, making it difficult to match coating colors efficiently and accurately.

Method used

The coating color is measured using a spectrophotometer. Combined with physical information machine learning algorithms and radiation transfer functions, the solution space is constrained by Kubelka-Monk theory to generate a set of color components. The toner concentration is then optimized using a machine learning model to match the target color.

Benefits of technology

It achieves more efficient and accurate coating color matching, reduces resource and time consumption, and improves the efficiency and accuracy of color matching.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118922696A_ABST
    Figure CN118922696A_ABST
Patent Text Reader

Abstract

A system may receive a color measurement of a first coating from a spectrophotometer, wherein the color measurement corresponds to a first color profile. Additionally, the system may process information derived from the first color curve using a physical information machine learning algorithm, where a radiation transfer function acts as a regularization agent that limits the space of allowable solutions within the physical information machine learning algorithm. The system may also generate a set of one or more color components corresponding to the first color curve in the event of mixing from an output of the physical information machine learning algorithm.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit and priority of U.S. Provisional Patent Application Serial No. 63 / 362,056, filed on March 29, 2022, and entitled “TECHNIQUES FOR COLOR MATCHING”, which is expressly incorporated herein by reference in its entirety. Additionally, this application claims the benefit and priority of U.S. Provisional Patent Application Serial No. 63 / 476,008, filed on December 19, 2022, and entitled “TECHNIQUES FOR COLOR MATCHING”, which is expressly incorporated herein by reference in its entirety. Technical Field

[0003] The present disclosure relates generally to coatings, and more particularly to techniques for color matching. Background Art

[0004] Modern coatings provide a variety of functions in industry and society. For example, vehicles and other types of machinery may be coated with paint or various other coatings to protect metal components from the elements (e.g., from rust) or to provide an aesthetically pleasing visual effect. The properties of the coating (e.g., color, visual effect, texture, etc.) may be determined in part based on the chemical composition of the coating. In some cases, a coating may be prepared to match an existing coating of the same color. Color matching techniques may be inefficient, time consuming, or inaccurate. Therefore, there are many opportunities for new methods and systems to improve the technology for color matching coatings. Summary of the invention

[0005] The disclosed system may include one or more processors and one or more computer-readable media. The computer-readable medium may have executable instructions stored thereon, which, when executed by the one or more processors, configure the computer system to perform various actions. The system may receive color measurements of a first coating from a spectrophotometer, wherein the color measurements correspond to a first color curve. Additionally, the system may process information obtained from the first color curve using a physical information machine learning algorithm, wherein a radiative transfer function acts as a regularization agent that limits the space of admissible solutions within the physical information machine learning algorithm. The system may also generate a set of one or more color components corresponding to the first color curve in the case of mixing based on the output of the physical information machine learning algorithm.

[0006] Additionally or alternatively, the disclosed computer-implemented method may include receiving color measurements of the first coating from a spectrophotometer, wherein the color measurements correspond to a first color curve. Additionally, the computer-implemented method may include processing information derived from the first color curve using a physical information machine learning algorithm, wherein a radiative transfer function acts as a regularization agent that restricts the space of admissible solutions within the physical information machine learning algorithm. Further, the computer-implemented method may include generating a set of one or more color components corresponding to the first color curve in the case of mixing based on the output of the physical information machine learning algorithm.

[0007] Additionally or alternatively, the disclosed non-transitory computer-readable medium may include one or more computer-readable storage media having executable instructions stored thereon, which when executed at a processor cause a computer system to perform a method for color matching. The method may include receiving a color measurement of a first coating from a spectrophotometer, wherein the color measurement corresponds to a first color curve. Additionally, the method may include processing information derived from the first color curve using a physical information machine learning algorithm, wherein a radiative transfer function acts as a regularization agent that restricts the space of admissible solutions within the physical information machine learning algorithm. Further, the method may include generating a set of one or more color components corresponding to the first color curve in the case of mixing based on the output of the physical information machine learning algorithm.

[0008] Additional features and advantages of the exemplary embodiments of the present disclosure will be set forth in the following description, and in part will be apparent in the description, or may be understood by the practice of the exemplary embodiments. The features and advantages of such embodiments may be realized and obtained by means of the instruments and combinations specifically indicated in the appended claims. These and other features will become more fully apparent from the following description and the appended claims, or may be understood by the practice of such exemplary embodiments as set forth below. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to describe the manner in which the above and other advantages and features of the present disclosure can be obtained, a more specific description of the present disclosure briefly described above will be presented by reference to its specific embodiments shown in the accompanying drawings. It should be understood that these drawings only depict typical embodiments of the present disclosure and are therefore not to be considered as limiting the scope thereof, and the present disclosure will be described and explained in more detail and in detail by using the drawings described below.

[0010] Figure 1 A schematic diagram showing a system for matching colors.

[0011] Figure 2 A schematic diagram showing a computer system for executing instructions.

[0012] Figure 3 A flow chart showing a system for color matching.

[0013] Figure 4 A flowchart showing the various steps used for color matching.

[0014] Figure 5 A flow chart showing one step for color matching.

[0015] Figure 6 A flowchart showing another step for color matching. DETAILED DESCRIPTION

[0016] The disclosed system, method or device is configured to perform color matching. The coating provides various functions, such as for protection or providing aesthetic visual effects. The coating can be expressed as a formula (e.g., a color formula or a chemical formula), which can indicate the color component. As used herein, "color component" includes chemical components within the coating formula. Examples of color components include toners, color pigments, effect pigments, base coatings, stains, dyes, etc. The method for determining the color formula may include using a spectrophotometer to measure a physical panel (e.g., sprayed panel, automotive panel, aircraft panel, hull, building wall, etc.) coated with the coating. The method may include using the measurement results to determine the formula (formula / recipe) of the coating. The method may include creating a coating using a determined formula, and repeatedly spraying and adjusting the formula to produce a coating that matches the coating of the physical panel. In some cases, the method for determining the color formula may be inefficient, inaccurate, or may require a large amount of resources or time. Therefore, it may be beneficial to implement the technology described herein for color matching.

[0017] A system (e.g., a computer system) for color matching can be implemented by a system including one or more processors and a computer-readable medium. For example, a computer-readable medium can have executable instructions stored thereon that, when executed, configure the system to identify a coating color. For example, a computer system can receive a color measurement of a first coating; the measurement is performed using a spectrophotometer. The coating can include at least one color property and at least one effect property. For example, a coating can be associated with a color and one or more visual effects (e.g., metallic, non-metallic, matte, etc.); however, some coatings may only include color features and may not include any visual effects. Measuring the coating using a spectrophotometer may produce a collection of spectral data representing the coating (e.g., optical properties).

[0018] The disclosed system can receive color measurements of a first coating from a spectrophotometer, wherein the color measurements correspond to a first color curve. Additionally, the system can process information obtained from the first color curve using a physical information machine learning algorithm, wherein a radiation transfer function acts as a regularization agent that limits the space of admissible solutions within the physical information machine learning algorithm. The system can also generate a set of one or more color components corresponding to the first color curve in the case of mixing based on the output of the physical information machine learning algorithm. The resulting set of one or more color components can be displayed to a user on a user interface, or sent directly to a blending machine for creating a coating. Therefore, in some configurations, the physical information machine learning algorithm is capable of generating the proposed color components based on a single color curve from a target coating.

[0019] As used herein, a "physical information machine learning algorithm" includes a machine learning algorithm embedded with knowledge of one or more physical laws that control the data set received by the machine learning algorithm. As used herein, a "radiative transfer function" includes a mathematical function related to the physical phenomenon of radiation propagation incident on a medium. For example, a radiative transfer function may include an algorithm derived from Kubelka-Munk theory related to the appearance of a coating on a substrate. Additionally, as used herein, a "regularization agent" includes one or more physical laws that are used within a physical information machine learning algorithm to limit the space of admissible solutions within and / or from the physical information machine learning algorithm. For example, the Kubelka-Munk algorithm can analyze a color coating by providing a range of one or more valid responses and / or intermediate results (also referred to herein as a "space"), thereby acting as a regularization agent. Therefore, the Kubelka-Munk algorithm can be embedded within a physical information machine learning algorithm to help the physical information machine learning algorithm calculate acceptable answers and / or prevent the physical information machine learning algorithm from providing physically impossible answers.

[0020] In some configurations, the system can further use a radiation transfer function (such as a Kubelka-Munk function) to determine a second color curve of the first coating based at least in part on the first color curve. As used herein, the Kubelka-Munk function can include any algorithm derived based on the principles of the Kubelka-Munk theory. The system can process the first color curve and the second color curve using a physical information machine learning algorithm. Additionally, the system can generate a third color curve based on the first color curve and the second color curve according to the physical information machine learning algorithm.

[0021] In addition, generating a set of one or more color components corresponding to the first color curve in the case of mixing according to the physical information machine learning algorithm may include identifying a set of one or more color components corresponding to the third color curve in the case of mixing. Additionally, generating the set of one or more color components may include outputting a color formula for the first coating, the color formula including the set of one or more color components associated with the third color curve.

[0022] Additionally or alternatively, identifying the set of one or more components can include selecting a subset of random components and incrementally adjusting the subset until the second color curve meets an accuracy threshold relative to the first color curve. For example, the system can calculate reflectance values ​​according to the Kubelka-Munk color theory, and the reflectance values ​​can be input into the machine learning model to adjust the subset of random components. Similarly, the system can incrementally adjust the concentrations associated with the one or more components so that the second color curve is based on the adjusted concentrations.

[0023] In some embodiments, the system can convert the color measurement data into three-dimensional color coordinates (e.g., in the CIELab color space, including lightness values ​​(L * ), red / green value (a * ) and blue / yellow value (b * )). This conversion can utilize a variety of different conversion methods known in the art, including but not limited to utilizing a spectrophotometer that automatically provides CIELab values, CIE color matching functions, lookup tables, or sums of Gaussian functions. Thus, the system can select components that surround the first coating in color space (e.g., having coordinates that satisfy a distance threshold relative to coordinates associated with color measurements). In some embodiments, the machine learning model can use a genetic algorithm, a gradient descent optimization process, a reinforcement learning process, a differential evolution optimization process, or any combination thereof.

[0024] The system can output (e.g., display to a user, send to a user computing device, etc.) a color formula comprising a set of components. The system can implement a set of operating rules that limit the number of effect pigments or components associated with the color formula. For example, a color formula can be limited to ten components such that it contains less than or equal to ten components or effect pigments. The system can add an indication of data associated with the first coating (e.g., color measurements) and the color formula to a data set (e.g., a historical coating panel data set) used to train the machine learning model.

[0025] Various aspects of the present disclosure may be implemented to achieve one or more potential advantages. For example, a system may use a hierarchical machine learning model in which the Kubelkamunk theory is used as a first approximation of the reflectivity curve, and then estimated reflectivity values ​​obtained from a historical set of coated panels are used as inputs to the machine learning model. Such machine learning models may enable more efficient, more accurate, or less time-consuming color matching procedures, among other benefits.

[0026] Other aspects of the disclosure, including examples, advantages, etc., will be described in the context of schematic diagrams, system diagrams, and method flows. As used throughout this application, unless otherwise stated, a single aspect of the invention disclosed and claimed encompasses multiple aspects, and vice versa.

[0027] Figure 1 A schematic diagram 100 of a system for matching colors is shown. Schematic diagram 100 may include a computer system 105 associated with a machine learning model 110 (e.g., stored on or in communication with the computer system 105). Schematic diagram 100 shows a spectrophotometer 115 and a user computing device 130 that share information or otherwise communicate with the computer system 105. Schematic diagram 100 may include a computer system 105 that may share data or perform Figure 1 Schematic diagram 100 may be used as part of a color matching process.

[0028] Color matching can be performed if an entity (e.g., a third party, user, auto repair shop, construction company, manufacturer, etc.) requests a coating that matches a pre-existing coating. For example, an auto repair shop may request a matching coating if it repairs a car that has a previously applied coating (e.g., due to collision damage) but does not have the coating or the formula / recipe for the coating. The requested coating may contain many different components or color pigments as well as effect pigments that provide a visual effect (e.g., metallic, non-metallic, matte, glossy, etc.).

[0029] In order to provide a matching coating or corresponding color formula, the first party may use a spectrophotometer to measure a previously applied coating (e.g., which may be located on a physical panel or vehicle). Additionally or alternatively, an auto repair shop may use a spectrophotometer to measure and send the resulting measurement data to the first party for processing. As described herein, the computer system 105 may implement a machine learning model 110, which may be a physical information machine learning algorithm. The machine learning model 110 may have been trained using a historical spray panel data set with known color curves and formulas. Therefore, the machine learning model 110 may be able to determine a color curve and a corresponding toner concentration or formula using, for example, spectral data measurements obtained using a spectrophotometer 115.

[0030] Computer system 105 may receive measurements 120 obtained using spectrophotometer 115. Measurements 120 may be associated with a first coating that includes at least one color property and at least one effect property (e.g., includes at least one type of toner). Measurements 120 may be examples of optical characteristics of the first coating represented as spectral data output from spectrophotometer 115. In some examples, the first coating may be associated with a previously coated physical panel or article (e.g., a vehicle).

[0031] The computer system 105 can determine a first color curve for the first coating based on the measurement results 120. The first color curve can be an example of a reflectance curve or spectrum representing the first coating (e.g., color components present in the first coating). Using the radiative transfer function, the computer system 105 can identify a second set of color components corresponding to the second color curve. Therefore, the second color curve is based on the first color curve through the radiative transfer function. Then, both the first color curve and the second color curve can be processed by the machine learning model 110 (e.g., a physical information machine learning algorithm).

[0032] In some examples, the computer system 105 may calculate a reflectance curve for each toner (e.g., using a machine learning model 110) so that a set of components is identified based on the calculated reflectance curve. For example, the computer system 105 may apply a Gaussian process regression model to data associated with each toner (e.g., spectral data for each toner) to determine the reflectance curve. In some embodiments, the computer system 105 may select a subset of random components and calculate the reflectance value of each toner in the subset of random components according to the Kubelka-Munk color theory. The computer system 105 may use the reflectance values ​​as input to the machine learning model 110, which may incrementally adjust the subset of random components until the resulting second color curve meets an accuracy threshold relative to the first color curve. Adjusting the subset of random components may include selecting different components or adjusting toner concentrations. In some embodiments, the computer system 105 may determine multiple color curves in parallel based on analyzing multiple components, toner concentrations, combinations of components, etc. For example, the computer system 105 may use different processors or other techniques to calculate multiple color curves in parallel. Similarly, the multiple components, toner concentrations, combinations of components, etc. used by the computer system 105 can be based on historical coating data used to train the machine learning model 110. In some configurations, the machine learning model 110 can use a genetic algorithm, a gradient descent optimization process, a reinforcement learning process, or a differential evolution optimization process.

[0033] Computer system 105 can convert data (e.g., spectral data) associated with color measurement 120 into three-dimensional coordinates in a color space (e.g., CIELab color space). Thus, computer system 105 can select components to test based on the coordinates in the color space. For example, computer system 105 can select components that meet a distance threshold relative to the coordinates of measurement 120 (e.g., components that surround the coordinates of measurement 120 in the color space).

[0034] The computer system 105 can output the set of components that match the measurement results 120 as a recipe 125 (e.g., a color recipe / formula) to the user computing device 130. The recipe 125 can indicate the components and the associated toner concentrations, as well as identifier information associated with the resulting coating. In some examples, the computer system 105 can implement a set of operating rules that include restrictions on the number of effect pigments or components associated with the recipe 125. For example, the computer system 105 can limit the recipe 125 to less than or equal to ten components or effect pigments. In some embodiments, the computer system 105 can add an indication of the data associated with the first coating and the color recipe to the data set used to train the machine learning model, so that the machine learning model 110 can develop a higher level of efficiency or accuracy when performing more color matching procedures. Implementing aspects of the schematic 100 can achieve more efficient, more accurate, or less time-consuming color matching procedures and other benefits.

[0035] Figure 2 An example of a computer system 200 that supports techniques for color matching according to aspects of the present disclosure is shown. The computer system 200 may implement the Figure 1 Aspects of the schematic diagram 100 described. Computer system 200 may include one or more processors 205 (e.g., processor 205a, processor 205b, processor 205c, etc.), a computer readable medium 210, and a collection of executable instructions 215. One or more processors 205 may include a general purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, an application specific integrated circuit, a field programmable gate array, or any combination thereof. Computer system 200 may be an example of a means for performing the methods and functions of the present disclosure as described herein.

[0036] For example, the set of executable instructions 215 may include instructions that, when executed by one or more processors 205, may cause the computer system 200 to at least perform the following operations: receive color measurements of a first coating from a spectrophotometer, wherein the color measurements correspond to a first color curve; process information derived from the first color curve using a physical information machine learning algorithm, wherein a radiative transfer function acts as a regularization agent that restricts the space of admissible solutions within the physical information machine learning algorithm; and generate a set of one or more color components corresponding to the first color curve in the case of a mixture based on the output of the physical information machine learning algorithm. Implementing aspects of the system 200 may achieve greater efficiency or greater accuracy of the color matching process, among other benefits.

[0037] Figure 3An example of a system 300 supporting techniques for color matching according to aspects of the present disclosure is shown. In some examples, aspects of the system 300 may implement aspects of the schematic diagram 100 or the system 200, as described with reference to FIG. Figure 1 and 2 For example, as described herein, the system 300 can be implemented by or otherwise associated with one or more processors or one or more computer-readable media. The system 300 can include a measurement result receiver 305, a color curve component 310, a toner identifier 315, an output component 320, or any combination thereof. The components of the system 300 as described herein can be as described herein. Figure 3 The system 300 may be configured to communicate or otherwise share information as shown or in other ways (such as between components not shown). In some examples, the system 300 may include additional components, or components may be removed or changed. The system 300 may be an example of a component for performing the methods and functions of the present disclosure as described herein.

[0038] The measurement receiver 305 may include a network connection, a wireless connection, a wired connection, or some other conventional computer I / O interface for communicating between the spectrophotometer and the computer. The received measurement 305 may receive a color measurement of the first coating from the spectrophotometer. The color measurement corresponds to the first color curve.

[0039] The toner identifier 315 may include a processing unit, such as those described herein, which are generally used for processing digital data and / or analog data. The toner identifier 315 uses a physical information machine learning algorithm to process information obtained from the first color curve. Additionally or alternatively, the identified toner 315 can use a physical information machine learning algorithm to process both the first color curve and the second color curve. The color curve component 310 can optionally use a radiation transfer function to calculate the second color curve based on the first color curve. Additionally, the radiation transfer function acts as a regularization agent that limits the space of admissible solutions within the physical information machine learning algorithm.

[0040] Toner identifier 315 can also apply Gaussian process regression model to the data associated with each toner in the second color curve, wherein the calculation reflectance curve is based at least in part on the application of Gaussian process regression model. In some instances, toner identifier 315 can convert the data associated with the color measurement result into three-dimensional coordinates in color space (e.g., CIELab color space), and can select one or more color components with coordinates around the first coating in color space based on the color measurement result of conversion first coating. In some instances, toner identifier 315 can determine multiple color curves in parallel based on analyzing multiple color components (comprising toner concentration, combination of components or any combination thereof). In some embodiments, analyzing multiple color components can be based on historical coating spraying.

[0041] In some configurations, the toner identifier 315 may select a subset of random components and incrementally adjust each subset of the random components until the second color curve meets an accuracy threshold relative to the first color curve. For example, the toner identifier 315 may calculate reflectance values ​​according to the Kubelka-Munk color theory. The toner identifier 315 may input the reflectance values ​​into a machine learning model to incrementally adjust each subset of the random components until the second color curve meets an accuracy threshold relative to the first color curve. The toner identifier 315 may incrementally adjust concentrations associated with one or more components, wherein the second color curve is based at least in part on adjusting the concentrations. The output component 320 may output a color formula for the first coating, the color formula comprising a set of one or more components.

[0042] The machine learning model can be a hierarchical machine learning model in which the Kubelka-Munk color theory is used as a first approximation of the reflectance curve, and then the estimated reflectance value obtained from the historical set of coated panels is used as the input of the machine learning model. The machine learning model can use a genetic algorithm. Similarly, the machine learning model can use a gradient descent optimization process, a reinforcement learning process, a differential evolution optimization process, or any combination thereof. The system 300 can implement a set of operating rules that include restrictions on the number of effect pigments associated with a color formula. For example, a color formula can include less than or equal to ten components or effect pigments. The various aspects of implementing the system 300 can achieve higher efficiency or higher accuracy of the color matching process and other benefits.

[0043] Figure 4 An example of a method flow 400 supporting a technique for color matching is shown. In some examples, the method flow 400 can be implemented by or otherwise associated with the schematic diagram 100, the system 200, or the system 300, as described in reference to FIG. Figure 1-3For example, the method flow 400 may correspond to executable instructions stored on a computer-readable medium, which configure a system (e.g., a computer system having one or more processors) to perform various actions. The steps shown in the method flow 400 may be performed in an order different from the order shown, or steps may be added or removed.

[0044] At 405 , method flow 400 can include receiving color measurements of a first coating from a spectrophotometer, wherein the color measurements correspond to a first color curve.

[0045] At 410 , method flow 400 may include processing information derived from the first color curve using a physics-informed machine learning algorithm, wherein the radiative transfer function acts as a regularization proxy that constrains the space of admissible solutions within the physics-informed machine learning algorithm.

[0046] At 415, method flow 400 can include generating a set of one or more color components corresponding to the first color curve in the mixed state based on the output of the physical information machine learning algorithm. Thus, the physical information machine learning algorithm can generate a coating recipe based on the reading of the first coating.

[0047] In some embodiments, the method flow 400 may include additional steps that are not shown. For example, the method flow 400 may include determining a second color curve for the first coating based at least in part on the first color curve. The method flow may further include processing the first color curve and the second color curve using a physical information machine learning algorithm. Further, the method flow may include generating a third color curve based on the first color curve and the second color curve according to the physical information machine learning algorithm. The third color curve is used by the physical information machine learning algorithm to generate one or more color components that correspond to the first color curve in the case of mixing.

[0048] In some instances, identifying a set of one or more components includes selecting a subset of random components and incrementally adjusting each subset of the random components until the second color curve meets an accuracy threshold relative to the first color curve. In some embodiments, adjusting each subset may include calculating reflectance values ​​according to the Kubelka-Munk color theory. Method flow 400 may include including the reflectance values ​​as input into a machine learning model to incrementally adjust each subset of the random components until the second color curve meets the accuracy threshold relative to the first color curve. In some instances, identifying a set of one or more components includes incrementally adjusting concentrations associated with the one or more components, wherein the second color curve is based on the adjusted concentrations.

[0049] In some embodiments, the machine learning model can use a genetic algorithm. Similarly, the machine learning model can use a gradient descent optimization process, a reinforcement learning process, a differential evolution optimization process, or any combination thereof. In some instances, the machine learning model comprises a hierarchical machine learning model, in which the Kubelka-Munk color theory is used as a first approximation of the reflectivity curve, and then the estimated reflectivity value obtained from the historical set of the coated panel is used as the input of the machine learning model. In some instances, the method flow 400 may include adding an indication of the data associated with the first coating and the color formula to the data set for training the machine learning model. The various aspects of the implementation method flow 400 can achieve higher efficiency or higher accuracy and other benefits of the color matching process.

[0050] Figure 5 An example of a method flow 500 supporting techniques for color matching according to aspects of the present disclosure is shown. In some examples, the method flow 400 can be implemented by or otherwise associated with the schematic diagram 100, the system 200, the system 300, or the method flow 400, as described in reference to FIG. Figure 1-4 For example, the method flow 500 may correspond to executable instructions stored on a computer-readable medium, which configure a system (e.g., a computer system having one or more processors) to perform various actions. The steps shown in the method flow 500 may be performed in an order different from the order shown, or steps may be added or removed.

[0051] In some examples, method flow 500 may be performed corresponding to method flow 400, as shown in FIG. Figure 4 For example, according to steps 405 to 410 of method flow 400, method flow 500 may include receiving a color measurement result of a first coating, the measurement result being performed using a spectrophotometer, the first coating including at least one color property and at least one effect property. Method flow 510 may also include determining a first color curve of the first coating based on the color measurement result.

[0052] At 515, method flow 500 can include calculating a reflectance curve associated with the first coating layer, wherein identifying the set of one or more components is based on calculating the reflectance curve associated with the first coating layer. In some embodiments, calculating the reflectance curve can include applying a Gaussian process regression model to data associated with each toner, wherein calculating the reflectance curve is based on applying the Gaussian process regression model.

[0053] According to 420 of method flow 400, method flow 500 may also include outputting a color formula of the first coating, the color formula including a set of one or more components. Implementing aspects of method flow 500 may achieve higher efficiency or higher accuracy of the color matching process, among other benefits.

[0054] Figure 6 An example of a method flow 600 supporting techniques for color matching according to aspects of the present disclosure is shown. In some examples, the method flow 400 can be implemented by or otherwise associated with the schematic diagram 100, the system 200, the system 300, the method flow 400, or the method flow 500, as described in reference to FIG. Figure 1-5 For example, method flow 600 may correspond to executable instructions stored on a computer-readable medium, which configure a system (e.g., a computer system having one or more processors) to perform various actions. The steps shown in method flow 600 may be performed in an order different from the order shown, or steps may be added or removed.

[0055] In some examples, method flow 600 may be performed corresponding to method flow 400 or method flow 500, as shown in FIG. Figure 4 and 5 For example, according to steps 405 to 410 of method flow 400, method flow 600 may include receiving a color measurement result of a first coating, the measurement result being performed using a spectrophotometer, the first coating including at least one color property and at least one effect property. Method flow 610 may also include determining a first color curve of the first coating based on the color measurement result.

[0056] Method flow 600 may also include calculating a reflectance curve associated with each toner, wherein identifying the set of one or more components is based on calculating the reflectance curve associated with each toner. In some embodiments, calculating the reflectance curve may include applying a Gaussian process regression model to data associated with each toner, wherein calculating the reflectance curve is based on applying the Gaussian process regression model.

[0057] In some examples, at 615, method flow 600 may also include determining multiple color curves for each toner in parallel based on analyzing multiple components, toner concentrations, combinations of components, or any combination thereof. In such examples, identifying a set of one or more components is based on determining multiple color curves in parallel. In some embodiments, analyzing multiple components, toner concentrations, combinations of components, or any combination thereof is based on historical coating spraying. Thus, a first color curve may describe the entire coating and / or a first color curve may describe a single toner within a coating.

[0058] According to 420 of method flow 400, method flow 600 may also include outputting a color formula of the first coating, the color formula including a set of one or more components. Implementing aspects of method flow 600 may achieve higher efficiency or higher accuracy of the color matching process and other benefits.

[0059] Although the subject matter has been described in language specific to structural features and / or methodological acts, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the described features or the acts or the order of the acts described. Rather, the described features and acts are disclosed as example forms of implementing the claims.

[0060] The present disclosure may include or utilize a special or general-purpose computer system, which includes computer hardware, such as, for example, one or more processors and system memory, as discussed in more detail below. Embodiments within the scope of the present disclosure also include physical media and other computer-readable media for carrying or storing computer-executable instructions and / or data structures. Such computer-readable media can be any available media that can be accessed by a general or special-purpose computer system. The computer-readable medium that stores computer-executable instructions and / or data structures is a computer storage medium. The computer-readable medium that carries computer-executable instructions and / or data structures is a transmission medium. Therefore, by way of example and not limitation, embodiments of the present disclosure may include at least two distinct types of computer-readable media: a computer storage medium and a transmission medium.

[0061] Computer storage media (e.g., including non-transitory computer-readable media) are physical storage media that store computer-executable instructions and / or data structures. Physical storage media include computer hardware, such as RAM, ROM, EEPROM, solid-state drive ("SSD"), flash memory, phase-change memory ("PCM"), optical disk storage, magnetic disk storage or other magnetic storage devices or any other hardware storage devices that can be used to store program code in the form of computer-executable instructions or data structures and can be accessed and executed by a general-purpose or special-purpose computer system to implement the disclosed functions of the present disclosure.

[0062] Transmission media may include networks and / or data links that may be used to carry computer executable instructions or program codes in the form of data structures and may be accessed by general or special computer systems. "Network" is defined as one or more data links that enable electronic data to be transmitted between computer systems and / or modules and / or other electronic devices. When information is transferred or provided to a computer system via a network or another communication connection (hard wiring, wireless, or a combination of hard wiring or wireless), the computer system may view the connection as a transmission medium. The combination of the above items should also be included within the scope of computer readable media.

[0063] Further, upon reaching various computer system components, program code in the form of computer executable instructions or data structures may be automatically transferred from a transmission medium to a computer storage medium (or vice versa). For example, computer executable instructions or data structures received over a network or data link may be buffered in RAM within a network interface module (e.g., a "NIC") and then ultimately transferred to computer system RAM and / or less volatile computer storage media at the computer system. Thus, it should be understood that computer storage media may be included in computer system components that also (or even primarily) utilize transmission media.

[0064] Computer executable instructions include, for example, instructions and data, which, when executed at one or more processors, cause a general purpose computer system, a special purpose computer system, or a special purpose processing device to perform a function or group of functions. Computer executable instructions may be binary, intermediate format instructions, or even source code, such as assembly language.

[0065] Those skilled in the art will appreciate that the present disclosure can be practiced in a network computing environment with various types of computer system configurations (including personal computers, desktop computers, laptop computers, message processors, handheld devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, network PCs, microcomputers, mainframe computers, mobile phones, PDAs, tablet computers, pagers, routers, switches, etc.). The present disclosure can also be practiced in a distributed system environment, where local computing systems and remote computer systems linked by networks (through hard-wired data links, wireless data links, or through a combination of hard-wired data links and wireless data links) all perform tasks. Therefore, in a distributed system environment, a computer system can include multiple component computer systems. In a distributed system environment, program modules can be located in local and remote memory storage devices.

[0066] It will also be appreciated by those skilled in the art that the present disclosure may be practiced in a cloud computing environment. A cloud computing environment may be distributed, but this is not required. In the case of a distribution, a cloud computing environment may be globally distributed within an organization and / or have components owned across multiple organizations. In this specification and the following claims, "cloud computing" is defined as a model for implementing on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage devices, applications, and services). The definition of "cloud computing" is not limited to any of the other numerous advantages that can be obtained from such a model when properly deployed.

[0067] Cloud computing models can be composed of various characteristics, such as on-demand self-service, broad network access, resource pooling, rapid elasticity, and metered services. Cloud computing models can also appear in the form of various service models, such as software as a service ("SaaS"), platform as a service ("PaaS"), and infrastructure as a service ("IaaS"). Cloud computing models can also use different deployment models, such as private cloud, community cloud, public cloud, hybrid cloud, etc.

[0068] Some embodiments, such as a cloud computing environment, may include a system comprising one or more hosts, each of which is capable of running one or more virtual machines. During operation, the virtual machine emulates an operational computing system that supports an operating system and possibly one or more other applications. In some embodiments, each host includes a hypervisor that simulates the virtual resources of the virtual machine using physical resources abstracted from the perspective of the virtual machine. The hypervisor also provides appropriate separation between virtual machines. Therefore, from the perspective of any given virtual machine, the hypervisor provides an illusion that the virtual machine is interfacing with physical resources, even if the virtual machine is only interfacing with the appearance of physical resources (e.g., virtual resources). Examples of physical resources include processing power, memory, disk space, network bandwidth, media drives, etc.

[0069] The present disclosure may be embodied in other specific forms without departing from the spirit or essential characteristics of the present invention. The described embodiments are considered to be merely illustrative and not restrictive in all respects. Therefore, the scope of the present disclosure is indicated by the appended claims rather than by the above description. All changes within the meaning and scope of the equivalent schemes of the claims will be covered within the scope of the claims.

Claims

1. A system for color matching, the system include: one or more processors; as well as one or more computer-readable media having stored thereon executable instructions that, when executed by the one or more processors, configure the system to: receiving color measurements of the first coating from a spectrophotometer, wherein the color measurements include a first color curve; processing information derived from the first color curve using a physically-informed machine learning algorithm, wherein a radiative transfer function acts as a regularization agent that constrains the space of admissible solutions within the physically-informed machine learning algorithm; and A set of one or more color components corresponding to the first coating layer in the mixed state is generated based on the output of the physical information machine learning algorithm.

2. The system of claim 1, wherein the one or more computer-readable media further contain executable instructions that, when executed at a processor, configure the system to: determining a second color curve for the first coating based at least in part on the first color curve; Processing the first color curve and the second color curve using the physical information machine learning algorithm; and and According to the physical information machine learning algorithm, a third color curve is generated based on the first color curve and the second color curve.

3. The system of claim 2, wherein the executable instructions for generating, from the physical information machine learning algorithm, the set of one or more color components corresponding to the first coating layer in the case of mixing include instructions that, when executed at a processor, configure the system to: identifying a set of the one or more color components that, when mixed, correspond to the third color curve; and A color formula for the first coating layer is output, the color formula including the set of the one or more color components.

4. The system of any one of claims 1 to 3, wherein the executable instructions for processing information derived from the first color curve using a physical information machine learning algorithm include instructions that when executed at a processor configure the system to: The first color curve and the second color curve are processed using a physics-informed machine learning algorithm, wherein the radiative transfer function serves as the regularization agent that constrains the space of admissible solutions within the physics-informed machine learning algorithm.

5. The system of any one of claims 1 to 4, wherein the executable instructions for generating the set of one or more color components include instructions that when executed at a processor configure the system to: A reflectance curve associated with each toner is calculated, wherein the reflectance curve is used to identify the set of one or more color components.

6. The system of claim 5, wherein the executable instructions for calculating the reflectivity curve include instructions that, when executed at a processor, configure the system to: A Gaussian process regression model is applied to data associated with each toner, wherein calculating the reflectance curve is based at least in part on applying the Gaussian process regression model.

7. The system of any one of claims 1 to 6, wherein the executable instructions for generating the set of one or more color components include instructions that when executed at a processor configure the system to: converting data associated with the color measurements into three-dimensional coordinates in a color space; and One or more components having coordinates in the color space surrounding the first coating layer are selected based at least in part on converting the color measurement of the first coating layer.

8. The system of claim 7, wherein the color space comprises a CIELab color space.

9. The system of any one of claims 1 to 8, wherein the executable instructions for generating the set of one or more color components include instructions that when executed at a processor configure the system to: A plurality of color curves are determined in parallel based at least in part on analyzing a plurality of components, toner concentrations, combinations of components, or any combination thereof.

10. The system of claim 9, wherein analyzing the plurality of components, toner concentrations, combinations of components, or any combination thereof is based at least in part on historical coating spraying.

11. The system of any one of claims 1 to 10, wherein the physical information machine learning algorithm comprises a machine learning algorithm that uses Kubelka-Munk theory as the regularization agent.

12. The system according to any one of claims 1 to 11, wherein a Kubelka-Munk function is used as the radiation transfer function to determine the first color curve and / or the second color curve.

13. The system of any one of claims 1 to 12, wherein said processing said information obtained from said first color curve using a physical information machine learning algorithm comprises: determining a second color curve for the first coating based at least in part on the first color curve, and generating a third color curve based on the first color curve and the second color curve according to the physical information machine learning algorithm, and Wherein generating a set of one or more color components corresponding to the third color curve rather than the first color curve when mixed based on the output of the physical information machine learning algorithm further includes outputting a color formula for the first coating, the color formula including the set of one or more color components.

14. The system of any one of claims 1 to 13, wherein the first color curve is a reflectance curve or spectrum representing the first coating.

15. The system of claim 14, wherein identifying the set of components is based on the calculating the reflectivity curve.

16. The system according to any one of claims 1 to 15, wherein generating a set of one or more color components according to the output of the physical information machine learning algorithm further include: Select a subset of random components; and incrementally adjusting each subset of the random components until the second color curve satisfies an accuracy threshold relative to the first color curve 17. The system of any one of claims 1 to 16, wherein the machine learning model is a hierarchical machine learning model in which the Kubelka-Munk color theory is used as a first approximation of the reflectance curve and then estimated reflectance values ​​obtained from a historical collection of coated panels are used as input to the machine learning model.

18. A computer-implemented method, executed on one or more processors, for color matching, the computer-implemented method include: receiving color measurements of the first coating from a spectrophotometer, wherein the color measurements include a first color curve; processing information derived from the first color curve using a physically-informed machine learning algorithm, wherein a radiative transfer function acts as a regularization agent that constrains the space of admissible solutions within the physically-informed machine learning algorithm; as well as A set of one or more color components corresponding to the first coating layer in the mixed state is generated based on the output of the physical information machine learning algorithm.

19. The computer-implemented method of claim 18, further comprising: include: determining a second color curve for the first coating based at least in part on the first color curve; Processing the first color curve and the second color curve using the physical information machine learning algorithm; and According to the physical information machine learning algorithm, a third color curve is generated based on the first color curve and the second color curve.

20. The computer-implemented method of claim 19, wherein the set of one or more color components corresponding to the first coating layer in the case of mixing is generated based on the physical information machine learning algorithm include: identifying a set of the one or more color components that, when mixed, correspond to the third color curve; as well as A color formula for the first coating layer is output, the color formula including the set of the one or more color components.

21. The computer-implemented method of any one of claims 18 to 20, wherein generating the set of one or more color components further comprises: include: A reflectance curve associated with each toner is calculated, wherein identifying the set of one or more color components is based at least in part on calculating the reflectance curve associated with each toner.

22. The computer-implemented method of any one of claims 18 to 21, wherein the reflectivity curve is calculated include: A Gaussian process regression model is applied to data associated with each toner, wherein calculating the reflectance curve is based at least in part on applying the Gaussian process regression model.

23. The computer-implemented method of any one of claims 18 to 21, wherein the set of one or more color components is generated include: converting data associated with the color measurements into three-dimensional coordinates in a color space; as well as One or more components having coordinates in the color space surrounding the first coating layer are selected based at least in part on converting the color measurement of the first coating layer.

24. The computer-implemented method of claim 23, wherein the color space comprises a CIELab color space.

25. The computer-implemented method of any one of claims 18 to 24, wherein the set of one or more color components is generated include: A plurality of color curves are determined in parallel based at least in part on analyzing a plurality of components, toner concentrations, combinations of components, or any combination thereof.

26. The computer-implemented method of any one of claims 18 to 25, wherein analyzing a plurality of components, toner concentrations, combinations of components, or any combination thereof is based at least in part on historical coating applications.

27. A non-transitory computer-readable medium comprising one or more computer-readable storage media having executable instructions stored thereon, the executable instructions, when executed at a processor, causing a computer system to perform a method for color matching, the method include: receiving a color measurement of the first coating from a spectrophotometer, wherein the color measurement corresponds to a first color curve; processing information derived from the first color curve using a physically-informed machine learning algorithm, wherein a radiative transfer function acts as a regularization agent that constrains the space of admissible solutions within the physically-informed machine learning algorithm; as well as A set of one or more color components corresponding to the first coating layer in the mixed state is generated based on the output of the physical information machine learning algorithm.