TWI895858B - Printing system and printing method with intelligent color-matching - Google Patents

Abstract

Provided is an intelligent color-matching printing system, including a storage device, a printing device, and a processing device. The storage device stores a program, a color-matching dataset, and a color-matching model. The color-matching dataset contains multiple pieces of color-matching data, and each piece of color-matching data includes an ink formula and its corresponding printing chromatic value. The printing device is controlled to produce printed products. The processing device loads the program from the storage device to execute: using the color-matching dataset to train the color-matching model; inputting the expected chromatic value into the trained color-matching model, and obtaining the predicted formula output by the trained color-matching model; and controlling, based on predicted formula, the printing device to output the printed product.

Description

Intelligent color matching printing system and printing method

The present invention relates to color printing technology, and more particularly to an intelligent color matching printing system and printing method.

A printing press is a mechanical device that prints text and/or images onto a substrate (also known as a "base material"). Because each substrate has its own unique color rendering, even with the same ink formulation, printed products on different substrates will have color differences. Furthermore, in the context of digital printing, the imaging principles of computer monitors and printing presses differ. Computer monitors typically use RGB color light, while printing presses typically use CMYK four-color inks for printing. These factors can cause the color of the printed product to differ from customer expectations, necessitating the use of dedicated personnel's experience in color formulation and adjustment. However, this manual color matching method still presents many unresolved issues. For example, while R2R (roll-to-roll) printing offers the advantages of large-scale, fast, continuous printing, and low costs, it relies on manual comparison for color matching and adjustment. This results in lengthy ink preparation times, high ink raw material consumption, and poor control over the color values of the finished print, hindering quality control.

As the manufacturing industry continues to invest in automation and digital production, an intelligent color matching solution is needed to improve printing system efficiency and batch stability.

An embodiment of the present invention provides an intelligent color matching printing system comprising a storage device, a printing device, and a processing device. The storage device stores a program, a color matching data set, and a color matching model. The color matching data set includes multiple color matching data, each of which includes an ink formula and its corresponding printing color value. The printing device is controlled to produce printed products. The processing device is connected to the printing device and loads the program from the storage device to execute the following: training the color matching model using the color matching data set; inputting expected color values into the trained color matching model to obtain a predicted formula output by the trained color matching model; and controlling the printing device to produce printed products based on the predicted formula.

In one embodiment, the ink formulation and the predicted formulation are related to the composition of multiple chemical materials, one of which is a surface tension adjusting agent.

In a further embodiment, the surface tension adjusting agent is present in the predicted formulation in an amount ranging from 0.1 to 3.5 phr (parts per hundreds of resin).

In a further embodiment, the surface tension adjusting agent has a spherical polyester structure having a hydroxyl group on the periphery and an extended chain composed of multiple alkyl groups.

In one embodiment, the color matching model is selected from a linear regression model, a neural network regression model, a decision tree regression model, and a random forest regression model.

Embodiments of the present invention further provide an intelligent color matching printing method. The printing method is executed by a processing device. The printing method includes the step of training a color matching model using a color matching data set. The printing method further includes the step of inputting expected color values into the trained color matching model to obtain a predicted formula output by the trained color matching model. The printing method further includes the step of controlling a printing device to produce a printed product based on the predicted formula.

The intelligent color matching printing system and printing method provided by this invention promotes the intelligence and efficiency of the printing process, effectively reducing the time required for the color matching process by more than 50% and saving more than 25% of ink raw materials.

10: Printing system

11: Processing device

12: Storage device

13: Printing device

14: Display device

15: Input device

200: Printing Method

S201-S203: Steps

20: Color matching model

21: Color matching dataset

211, 212, 21N: Color matching data

211A, 212A, 21NA: Ink formula

211B, 212B, 21NB: Printing color values

22: Expected color value

23: Predictive Recipe

13: Printing device

25: Printed Product

The present invention will be better understood from the following description of exemplary embodiments in conjunction with the accompanying drawings. It should also be understood that the order of execution of the blocks in the flowcharts may be changed, and/or certain blocks may be changed, deleted, or combined.

Figure 1 is a system block diagram of an intelligent color matching printing system according to one embodiment of the present invention.

Figure 2A is a flow chart of an intelligent color matching printing method according to one embodiment of the present invention.

Figure 2B is a schematic diagram of an intelligent color matching printing method according to one embodiment of the present invention.

The following description lists various embodiments of the present invention and is not intended to limit the scope of the present invention. The actual scope of the invention is defined by the scope of the patent application.

In the various embodiments listed below, the same reference numerals will be used to represent the same or similar elements or components.

In this specification and in the patent application, serial numbers, such as "first," "second," etc., are used for convenience only and do not have a sequential order.

The following descriptions of device or system embodiments also apply to method embodiments, and vice versa.

An embodiment of the present invention employs machine learning, whereby a color matching dataset is constructed by collecting and cleaning ink formulation and color data. This dataset is then used to train a color matching model. The trained color matching model can be used to predict the ideal color ink formulation, which can then be used to produce printed products. In a further embodiment, a novel surface tension adjustment agent is introduced into the ink formulation to enhance the convergence and accuracy of the color matching model, and improve print resolution and color uniformity.

Figure 1 is a system block diagram of an intelligent color matching printing system 10 according to one embodiment of the present invention. As shown in Figure 1, the printing system 10 includes a processing device 11, a storage device 12, and a printing device 13.

The processing device 11 may include one or more hardware components for executing instructions, such as a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor, a controller, a microcontroller, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a system on a chip (SoC), etc., but the present invention is not limited thereto. The processing device 11 loads a program from the storage device 12 to execute an intelligent color matching printing method. The details of this printing method will be described in detail below with reference to Figures 2A and 2B. In one embodiment, these hardware components can be deployed locally (i.e., at the on-site location of the printing device 13 that produces the finished printed product) to perform all steps of the aforementioned printing method. In another embodiment, these hardware components can be deployed remotely and locally to perform the steps of training the color matching model and using the color matching model to predict the ink formula, respectively.

Storage device 12 can be any device with non-volatile memory (e.g., read-only memory, electrically-erasable programmable read-only memory (EEPROM), flash memory, non-volatile random access memory (NVRAM)), such as a hard disk drive (HDD) array, a solid-state drive (SSD), or an optical disk, but the present invention is not limited thereto. Storage device 12 can be fully or partially deployed locally (i.e., at the site where printing device 13 that produces the printed product is located) or remotely, but the present invention is not limited thereto. In this embodiment of the present invention, storage device 12 stores a program containing multiple instructions for implementing the aforementioned printing method. When processing device 11 loads the program from storage device 12, it executes these instructions to implement the aforementioned printing method. Furthermore, storage device 12 stores color matching data sets and color matching models. Details of these color matching data sets and color matching models will be described later with reference to FIG. 2B.

Printing device 13 can be any color printer used to print text and/or images on a substrate. The substrate can be various types of paper (e.g., newsprint, letterpress, gravure, map paper, poster paper, etc.), various fiber fabrics (e.g., clothing, scarves, ties, bed sheets, quilts, pillowcases, etc.), various plastic rolls (e.g., films, sheets, artificial leather, wallpaper, etc.), ceramics, metals, etc., but the present invention is not limited thereto. The ink used in printing device 13 is typically mixed using the CMYK color separation model (although the present invention is not limited thereto), where C represents cyan, M represents magenta, Y represents yellow, and K represents black. In an embodiment of the present invention, the printing device 13 is connected to the processing device 11 in a wired or wireless manner and is controlled by the processing device 11 to produce printed products.

In one embodiment, the printing system 10 further includes a display device 14 and an input device 15. The display device 14 can be any type of display, such as an LCD display, an LED display, an OLED display, an electronic paper, a projection device, or a plasma display, but the present invention is not limited thereto. The input device 15 can be, for example, a mouse, a keyboard, a control panel (console), a touch display element, a voice input device, or a keypad, but the present invention is not limited thereto. In this embodiment, the display device 14 can present the formula predicted by the above-mentioned color matching model to the user (such as a colorist or colorist), and the printing system 10 allows the user to fine-tune the predicted formula through the input device 15. In addition, the printing system 10 also allows the user to input the expected color value of the printed product through the input device 15.

In one embodiment, the printing system 10 may further include a communication interface (not shown in FIG. 1 ), which allows the printing system 10 to communicate with other devices to obtain and/or update data required to implement the above-mentioned intelligent color matching printing method, such as the above-mentioned program, color matching data set, and color matching model. The communication interface can be a wired communication interface, such as High Definition Multimedia Interface (HDMI), DisplayPort (DP) interface, embedded DisplayPort (eDP) interface, Universal Serial Bus (USB) interface, USB Type-C interface, Thunderbolt interface, Digital Video Interface (DVI), and combinations thereof. It can also be a wireless communication interface, such as a 5G wireless system, Bluetooth, Wi-Fi, or Near Field Communication (NFC) interface, but this disclosure is not limited thereto.

Figure 2A is a flow chart of an intelligent color matching printing method 200 according to one embodiment of the present invention. As shown in Figure 2A , printing method 200 includes steps S201-S203. Accordingly, Figure 2B is a schematic diagram of printing method 200. Please refer to Figures 2A and 2B together for a clearer understanding of the embodiments of the present invention.

The printing method 200 can be executed by the processing device 11 in Figure 1. In step S201, the color matching model 20 is trained using the color matching dataset 21. Then, step S202 is performed.

In step S202, the expected color value 22 is input into the trained color matching model 20 to obtain the predicted formula 23 output by the trained color matching model 20. Then, step S203 is performed.

In step S203, based on the predicted recipe 23, the printing device 13 is controlled to produce the printed product 25.

As shown in Figure 2B , color matching data set 21 includes multiple color matching data 211-21N. Each color matching data entry includes an ink formula and its corresponding print color value. For example, color matching data 211 includes ink formula 211A and its corresponding print color value 211B, color matching data 212 includes ink formula 212A and its corresponding print color value 212B, color matching data 21N includes ink formula 21NA and its corresponding print color value 21NB, and so on. In other words, color matching data 211 is associated with the mapping between ink formula 211A and print color value 211B, color matching data 212 is associated with the mapping between ink formula 212A and print color value 212B, color matching data 21N is associated with the mapping between ink formula 21NA and print color value 21NB, and so on.

In one embodiment, the color matching data 211-21N in the color matching data set 21 can be collected through printing experiments on specific substrates. Different substrates can have their own corresponding color matching data sets 21.

In one embodiment, the printed color values 211B-21NB are measured using a colorimeter (also known as a "colorimeter" or "colorimeter") from printed products produced by the printing device 13 using the corresponding ink formulations 211A-21NA. The printed color values 211B-21NB are typically defined using the CIELAB color space (also known as "CIE L*a*b*") (although the present invention is not limited thereto), where L* represents lightness, a* represents green-red chromaticity, and b* represents green-yellow chromaticity.

In one embodiment, the color matching model 20 is a regression model, in which the dependent variable is the printed color value and the independent variable is the ink formula. In this embodiment, during the model training process (i.e., step S201), loss functions such as mean square error (MSE), mean absolute error (MAE), mean squared logarithmic error (MSLE), root mean square error (RMSE), mean absolute percentage error (MAPE), standard deviation (σ), normalized mean square error, absolute error, and relative error can be used to calculate the loss value representing the difference between the predicted recipe output by the regression model and the actual printed color value. For example, if ink formula 211A is input into the regression model to obtain a first predicted formula (not shown in FIG. 2B ), the first loss value represents the difference between the first predicted formula and the printed color value 211B. If ink formula 212A is input into the regression model to obtain a second predicted formula (not shown in FIG. 2B ), the second loss value represents the difference between the second predicted formula and the printed color value 212B, and so on. Furthermore, an optimizer can be used to recursively adjust the parameters of the regression model to minimize the loss value, thereby optimizing the regression model. The optimizer can be implemented using algorithms such as gradient descent (GD), batch gradient descent (BGD), stochastic gradient descent (SGD), mini-batch gradient descent (MBGD), momentum gradient descent (Momentum), Nesterov accelerated gradient descent (NAG), adaptive gradient algorithm (AdaGrad), RMSprop (root mean square prop), AdaDelta, or adaptive moment estimation (Adam), but the present invention is not limited thereto. For example, an optimizer using gradient descent can calculate the gradient by performing a partial derivative on the loss function. The gradient is then used to adjust the parameters of the regression model to reduce the loss. By repeatedly feeding back results and updating parameters, the loss value is gradually reduced until it converges to a minimum value.

In one embodiment, the color matching model 20 is selected from a variety of regression models based on their measured training error and testing error. Table 1 below shows the training results for linear regression, neural networks (with an optimizer using stochastic gradient descent), decision trees, and random forests, with the RMSE value used as the training error and testing error.

As can be seen from Table 1, the training errors of the linear regression model and the neural network regression model are high (both above 0.5), indicating that the models are underfitting. The decision tree regression model has a large gap between its training error (0.000) and its test error (1.019), indicating that the model is overfitting. Overfitted and underfitted regression models have relatively poor predictive capabilities. Therefore, in this embodiment, a random forest regression model is used as the color matching model 20.

In one embodiment, ink formulations 211A-21NA and predicted formulation 23 are associated with a composition of multiple chemical materials, one of which is a surface tension modifier. In other words, ink formulations 211A-21NA and predicted formulation 23 contain the surface tension modifier, and the strain coefficient of color matching model 20 includes the content of the surface tension modifier. The inclusion of the surface tension modifier can improve printing resolution, as well as color stability and uniformity.

In a further embodiment, the surface tension-adjusting agent comprises a spherical polyester structure with a hydroxyl group and an extended chain of multiple alkyl groups around the periphery. The spherical polyester structure can also be symmetrical. This polyester structure can further enhance surface tension, preventing ink collapse and over-wetting on the substrate, which can lead to blurred printed dots.

In a further embodiment, the surface tension modifier content in the predicted formulation 23 ranges from 0.1 to 3.5 phr (parts per hundreds of resin). Experimental results show that compared to a formulation without the surface tension modifier, the printed product 25 produced using a formulation containing 0.1 to 3.5 phr of the surface tension modifier exhibits a 10-25% improvement in print resolution and a 15-30% improvement in color stability (as measured by the color consistency index (CCI)). Furthermore, the average Delta E value of the printed product 25 (i.e., the square root of the sum of the squared differences in L*, a*, and b* between the printed product 25 and the expected color values 22) was significantly reduced from 6.39 to 3.89. Furthermore, the training error of the color matching model 20 was reduced from 0.349 to 0.302, and the testing error was reduced from 0.999 to 0.901. This demonstrates that the addition of the surface tension adjustment agent further improves the training convergence and color matching accuracy of the color matching model 20.

The printing method of the present invention, or a specific form or portion thereof, can be included in the form of program code on a physical medium, such as a floppy disk, an optical disk, a hard disk, or any other machine-readable (e.g., computer-readable) storage medium. When the program code is loaded and executed by a machine, such as a computer, the machine becomes a device or system for participating in the present invention. The printing method, system, and device of the present invention can also be transmitted in the form of program code via some transmission medium, such as wires or cables, optical fibers, wireless networks, satellite signals, or any other transmission type. When the program code is received, loaded, and executed by a machine, such as a computer, a handheld device, or a wearable device, the machine becomes a device or system for participating in the present invention. When implemented on a general-purpose processor, the code combines with the processor to provide a unique device that operates similarly to application-specific logic circuits.

The intelligent color matching printing system and printing method provided by this invention promotes the intelligence and efficiency of the printing process, effectively reducing the time required for the color matching process by more than 50% and saving more than 25% of ink raw materials.

The above paragraphs describe various aspects. Obviously, the teachings herein can be implemented in a variety of ways, and any specific architecture or functionality disclosed in the examples is merely representative. Based on the teachings herein, those skilled in the art will understand that each aspect disclosed herein can be implemented independently, or two or more aspects can be combined.

Although the present invention has been described above with reference to the embodiments, this is not intended to limit the present invention. Anyone skilled in the art may make modifications and improvements without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the invention shall be determined by the scope of the patent application attached hereto.

200: Printing Method

S201-S203: Steps

20: Color matching model

21: Color matching dataset

211, 212, 21N: Color matching data

211A, 212A, 21NA: Ink formula

211B, 212B, 21NB: Printing color values

22: Expected color value

23: Predictive Recipe

13: Printing device

25: Printed Product

Claims (10)

An intelligent color matching printing system comprises: A storage device storing a program, a color matching data set, and a color matching model, wherein the color matching data set includes multiple color matching data, each color matching data set including an ink formula and a printing color value corresponding to the ink formula; A printing device controlled to produce a printed product; and A processing device connected to the printing device and loading the program from the storage device to execute: Training the color matching model using the color matching data set; Inputting an expected color value into the trained color matching model to obtain a predicted formula output by the trained color matching model; and Controlling the printing device to produce the printed product based on the predicted formula. The printing system of claim 1, wherein the ink formula and the predicted formula are related to the composition of multiple chemical materials, one of which is a surface tension adjusting agent. The printing system of claim 2, wherein the content of the surface tension adjusting agent in the predicted formulation is between 0.1 and 3.5 phr (parts per hundreds of resin). The printing system of claim 2, wherein the surface tension adjusting agent has a spherical polyester structure, the outer periphery of the polyester structure has a hydroxyl group and an extended chain composed of multiple alkyl groups. The printing system of claim 1, wherein the color matching model is selected from a linear regression model, a neural network regression model, a decision tree regression model, and a random forest regression model. A printing method for intelligent color matching, performed by a processing device, includes: Training a color matching model using a color matching dataset, wherein the color matching dataset includes multiple color matching data entries, each of which includes an ink formula and a printing color value corresponding to the ink formula; Inputting an expected color value into the trained color matching model to obtain a predicted formula output by the trained color matching model; and Based on the predicted formula, controlling a printing device to produce a printed product. The printing method of claim 6, wherein the ink formula and the predicted formula are related to the composition of multiple chemical materials, one of which is a surface tension adjusting agent. The printing method of claim 7, wherein the content of the surface tension adjusting agent in the predicted formulation is between 0.1 and 3.5 phr (parts per hundreds of resin). The printing method of claim 7, wherein the surface tension adjusting agent has a spherical polyester structure, the outer periphery of the polyester structure has a plurality of hydroxyl groups and an extended chain composed of a plurality of alkyl groups. The printing method of claim 6, wherein the color matching model is selected from a linear regression model, a neural network regression model, a decision tree regression model, and a random forest regression model.