Printing quality improving method based on printing parameter optimization

Abstract

The invention relates to the technical field of printing, and discloses a printing parameter optimization-based printing quality improvement method, which comprises the following steps of: constructing a multi-dimensional color characteristic model through spectral data; collecting illumination conditions and related parameters of an observation environment in real time through a spectrum sensor, and generating an environment parameter data set; establishing an observer personalized visual perception model through a color matching experiment, wherein the personalized visual perception model maps the physical stimulation into perception response; dynamically adjusting printing parameters based on the multi-dimensional color characteristic model, the environmental parameter data set and the personalized visual perception model; according to the method, printing output is monitored in real time and printing parameters are corrected in the printing process to minimize perceptual printing differences, so that environmental factors and individual visual perceptual differences can be comprehensively considered, the printing parameters can be optimized in real time, and the printing quality and the color rendition precision are improved.

Description

Technical Field

[0001] The present invention relates to the field of printing technology, and more particularly to a method for improving printing quality based on printing parameter optimization. Background Art

[0002] With the continuous development of printing technology, the quality requirements for printed products are becoming increasingly stringent, especially in terms of color reproduction accuracy. However, in traditional printing processes, changes in environmental factors and individual differences in observer visual perception often lead to deviations between the color performance of the final printed product and the expected effect. Traditional color management technologies are mainly based on standard observer models and standard light source conditions, which cannot effectively cope with the dynamic changes in actual printing and observation environments. At the same time, traditional methods often ignore the individual differences in color perception among different observers, resulting in inconsistent visual effects of printed products under different observation conditions. In addition, the parameter optimization process in existing technologies is mostly offline static optimization, which is difficult to dynamically adjust according to real-time environmental conditions and observer characteristics.

[0003] Therefore, how to comprehensively consider environmental factors and individual visual perception differences and optimize printing parameters in real time to improve printing quality and color reproduction accuracy has become a technical problem that needs to be solved urgently. Summary of the Invention

[0004] The present invention provides a method for improving printing quality based on printing parameter optimization, which solves the technical problem in the prior art of being able to comprehensively consider environmental factors and individual visual perception differences and to optimize printing parameters in real time to improve printing quality and color reproduction accuracy.

[0005] The present invention provides a method for improving printing quality based on printing parameter optimization, comprising the following steps: Build a multidimensional color characteristic model using spectral data. The multidimensional color characteristic model includes training data on printing conditions, equipment characteristics, ink characteristics, and paper characteristics. Based on the construction of a multi-dimensional color characteristic model, the lighting conditions and related parameters of the observed environment are collected in real time through spectral sensors to generate an environmental parameter data set; Based on the environmental parameter dataset, a personalized visual perception model of the observer is established through color matching experiments. The personalized visual perception model maps physical stimuli to perceptual responses. Based on a multidimensional color characteristic model, an environmental parameter dataset, and a personalized visual perception model, an iterative color separation algorithm is used to obtain the optimal ink distribution. Color conversion parameters are adjusted in accordance with real-time environmental conditions, and color compensation parameters are fine-tuned using a personalized visual perception model to dynamically adjust printing parameters. Based on dynamic adjustment of printing parameters, the printing output is monitored in real time during the printing process and the printing parameters are corrected to minimize the perceived printing differences.

[0006] Furthermore, the step of constructing a multidimensional color characteristic model based on spectral data includes: collecting color output spectrum data of a printing device under different printing conditions, wherein the printing conditions include temperature, humidity, paper type, and ink type; Use piecewise polynomial interpolation algorithm to model spectral data and construct color space mapping relationship; A comprehensive color response model including device characteristics, ink characteristics and paper characteristics is constructed, wherein the comprehensive color response model is expressed as a product of the device response function, the ink response function and the paper response function.

[0007] Furthermore, the piecewise polynomial interpolation algorithm divides the wavelength range into multiple sub-intervals, applies a different polynomial function to each sub-interval, and obtains the polynomial coefficients by solving the least square method.

[0008] Furthermore, the step of using a spectral sensor to collect lighting conditions and related parameters of the observation environment in real time includes: Configure a micro-spectral sensor array system to measure real-time environmental spectral parameters; Collect environmental spectral distribution data, measure key parameters of ambient illumination and color temperature, and obtain chromaticity coordinates and color purity; Use distributed sensor arrays to capture spatial illumination unevenness and generate a spatial illumination distribution model; The collected environmental data is analyzed in time series to generate an environmental parameter change model to predict the short-term change trend of the ambient lighting conditions.

[0009] Furthermore, the micro spectral sensor array system includes a plurality of distributed spectral sensor units and a central data processing unit, each sensor unit captures spectral information within a wavelength range of 380nm to 780nm and transmits data to the central processing unit.

[0010] Furthermore, the step of establishing the observer's personalized visual perception model through the color matching experiment includes: Perform color matching experiments to obtain observer-specific color perception characteristics data; Apply neural coding theory to establish a mathematical mapping relationship from physical stimuli to perceptual responses; Combined with color appearance models to obtain color perception properties under specific conditions; Individual perception data is modeled through machine learning algorithms to train personalized visual perception models.

[0011] Furthermore, the personalized visual perception model adopts a multi-layer perceptron structure, including: An input layer that receives physical stimulus parameters and environmental conditions; Three hidden layers, containing 64, 128, and 64 neurons respectively, using a nonlinear activation function; An output layer that produces predicted individual perceptual responses, including perceived brightness, hue, saturation, and chromaticity values.

[0012] Furthermore, the dynamic adjustment of printing parameters includes: The iterative color separation algorithm is used to solve the optimization problem of the optimal ink volume combination; Obtain environmental compensation factors based on real-time environmental conditions and adjust color conversion parameters; Combined with the individualized visual perception model, the individualized compensation factor is obtained and the color compensation parameters are fine-tuned; Parameter settings are continuously optimized through a closed-loop feedback mechanism.

[0013] Furthermore, the step of real-time monitoring of printing output and correcting printing parameters includes: Configure the spectrum detection sensor system to monitor the printing output in real time; Obtain the difference between actual output and expected perception based on an individualized visual perception model; Apply real-time correction algorithms to automatically adjust printing parameters to minimize perceived differences; Establish an optimized data recording system to collect information on parameters, environmental conditions, correction operations and final quality evaluation during the printing process.

[0014] A computer-readable storage medium is used for storing computer-readable instructions. When the computer-readable instructions are read by a computer, a printing quality improvement method based on printing parameter optimization can be executed.

[0015] The beneficial effects of the present invention are as follows: the present invention constructs a multidimensional color characteristic model using spectral data, the multidimensional color characteristic model including training data on printing conditions, equipment characteristics, ink characteristics, and paper characteristics; based on the constructed multidimensional color characteristic model, a spectral sensor is used to collect lighting conditions and related parameters of the observation environment in real time to generate an environmental parameter dataset; based on the environmental parameter dataset, a personalized visual perception model of an observer is established through a color matching experiment, the personalized visual perception model mapping physical stimuli to perceptual responses; based on the multidimensional color characteristic model, the environmental parameter dataset, and the personalized visual perception model, printing parameters are dynamically adjusted; based on the dynamic adjustment of printing parameters, the printing output is monitored in real time during the printing process and the printing parameters are corrected to minimize perceived printing differences, thereby enabling the printing parameters to be dynamically adjusted according to the real-time environmental data, so that the printed product maintains a consistent visual effect under different environmental conditions; and parameters are optimized according to the visual characteristics of different observers to provide a personalized visual experience. Through a real-time monitoring and feedback mechanism, fluctuations and deviations in the printing process are reduced, and the stability of printed product quality is improved. The debugging time and scrap rate are reduced, and the overall efficiency of printing production is improved, ensuring that the printed product has a consistent perceptual experience across different observation environments and for different observers. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The present invention is a flowchart of a method for improving printing quality based on printing parameter optimization. DETAILED DESCRIPTION

[0017] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. Furthermore, features described in some examples may be combined in other examples.

[0018] At least one embodiment of the present invention discloses a method for improving printing quality based on printing parameter optimization, such as Figure 1 As shown, including: Step 1: Build a multidimensional color characteristic model using spectral data. The multidimensional color characteristic model includes training data on printing conditions, equipment characteristics, ink characteristics, and paper characteristics.

[0019] Apply the piecewise polynomial interpolation algorithm to process the color output spectrum data collected under printing conditions to generate a high-precision device color characteristic model. This step includes: Collect the color output spectrum data of the printing equipment under different printing conditions (including temperature, humidity, paper type, ink type, etc.). Specifically, use a spectrometer to measure the standard color card at 1nm intervals within the wavelength range of 380nm to 780nm to obtain the reflectance curve ,in Indicates wavelength; The piecewise polynomial interpolation algorithm is used to model the spectral data and construct the color space mapping relationship. The algorithm divides the wavelength range into subintervals, for each subinterval Apply different polynomial functions ,as follows: ; in, Item 0, Item 1, Item The coefficient of the term, is the total number of polynomial coefficients, is the polynomial order, which is solved by the least squares method; Build a comprehensive color response model that integrates device characteristics, ink characteristics, and paper characteristics into a unified color response function : ; in, is the device response function, is the ink response function, is the paper response function, Control parameters for the device.

[0020] Through the above steps, a high-precision multidimensional model is established that accurately describes how a printing device produces specific color output under different printing conditions. This model is the basis for subsequent environmental adaptation and personalized color perception optimization.

[0021] Step 2: Based on the construction of a multidimensional color characteristic model, the lighting conditions and related parameters of the observation environment are collected in real time through a spectral sensor to generate an environmental parameter dataset.

[0022] Use a micro-spectral sensor array to collect the lighting conditions and related parameters of the observation environment to generate an environmental parameter dataset. This step includes: A micro-spectral sensor array system is configured to measure real-time environmental spectral parameters. The system includes multiple distributed spectral sensor units and a central data processing unit, wherein each sensor unit can capture spectral information within the wavelength range of 380nm to 780nm and transmit the data to the central processing unit; Collect environmental spectral distribution data , measure ambient illumination , color temperature And other key parameters, and calculate the chromaticity coordinates And color purity: ; in, is a constant, is the visual brightness function, is the wavelength, is the color purity, Represents the differential term of wavelength, used to integrate wavelength; Use distributed sensor arrays to capture spatial illumination unevenness and generate spatial illumination distribution models ,in, Indicates spatial location. Records the observer's perspective and focus area data at the same time ; Perform time series analysis on the collected environmental data to generate an environmental parameter change model ,in Represents time and is used to predict short-term changing trends in ambient light conditions.

[0023] Through these steps, a complete environmental perception data acquisition system is formed, which acquires environmental parameters for color optimization in real time and provides a data foundation for subsequent dynamic parameter adjustments. This system is characterized by its ability to capture spatial distribution unevenness and temporal dynamic changes, overcoming the limitations of traditional color management systems that rely on single, static lighting conditions.

[0024] Step 3: Based on the environmental parameter dataset, a personalized visual perception model of the observer is established through a color matching experiment. The personalized visual perception model maps physical stimuli to perceptual responses.

[0025] Through neural coding theory and color matching experiments, a personalized visual perception model of the observer is constructed to generate a personalized color perception parameter dataset. This step includes: Design and execute color matching experiments to obtain observer-specific color perception characteristic data. Specifically, by having observers match standard color samples under different lighting conditions, record their subjective color matching results, and collect matching data sets containing attributes such as hue, saturation, and brightness. ,in, For standard samples, Matching results for observers, is the sample size; Neural coding theory is applied to establish a mathematical mapping relationship from physical stimulation to perceptual response. This mapping relationship is expressed as: ; in, Indicates the sensory response, 、 、 represent the response values ​​of the first, second and third cone cells, respectively. Indicates contextual factors (such as surrounding color, background brightness, etc.), Represents the time factor (such as visual adaptation state), f represents the neural encoding mapping function, which maps the first, second and third cone cell response values ​​​​(L, M, S), contextual factors (C) and time factors (t) to the final perceptual response R. This function is implemented by a neural network, which contains components such as weight matrix, bias term and nonlinear activation function.

[0026] ; in, and denote the first and second weight matrices respectively, and denote the first and second bias terms respectively, Represents a nonlinear activation function. The specific implementation of the neural encoding model includes the following steps: Construct a three-layer neural network structure, including an input layer, a hidden layer, and an output layer, where the input layer receives the first, second, and third cone cell response values 、 、 and contextual factors and time factors ; The hidden layer uses the ReLU activation function and has 128 neurons, processing the input through the following calculation: ; in, represents the hidden layer output, represents the rectified linear unit activation function, represents the third bias term; contextual factors and time factors Connected with the hidden layer output to form an enhanced feature representation : ; in, represents the hidden layer output, Represents contextual factors, Indicates the time factor, The output layer uses a linear activation function to map the enhanced features to the perceptual response space: ; in, represents enhanced feature representation, represents the first weight matrix, represents the fourth bias term, Indicates a perceptual response; The model is trained using the observer color matching experimental data and the stochastic gradient descent algorithm is used to minimize the mean square error between the predicted response and the actual response. : , in, is the sensory response predicted by the model, is the actual perceptual response obtained in the experiment, is the sample size.

[0027] In specific application scenarios, this neural encoding model can be used to predict a specific observer's perceptual response to printed colors under different environmental conditions. For example, when a printed product is moved from an office fluorescent lighting environment (color temperature approximately 6500K) to a home incandescent lighting environment (color temperature approximately 2700K), the model can accurately predict the different perceptual responses of the same observer to the same printed color, thereby guiding the adjustment of printing parameters to ensure perceptual consistency.

[0028] Combined with the CIECAM02 color appearance model, calculate the color perception attributes under specific conditions. ,tone and saturation The calculation formula is as follows: ; ; ; in, is the white point brightness, is the brightness adaptation factor, and are the first and second environmental parameters, and are the first and second chromaticity coordinates, is the temporary chroma, Environmental factors; Model individual perception data through machine learning algorithms and train personalized visual perception models The model uses a multi-layer perceptron structure, with physical stimulus parameters and environmental conditions as input and the output as the predicted individual perceptual response: ; in, For individual perception response, is the physical stimulus parameter, The specific implementation of the personalized visual perception model includes the following steps: Construct a five-layer fully connected neural network, including an input layer, three hidden layers, and an output layer; The input layer receives a 10-dimensional feature vector, including physical stimulus parameters (such as RGB / CMYK values, illuminance, wavelength distribution, etc.) and environmental condition parameters (such as lighting color temperature, ambient reflectivity, observation angle, etc.); The three hidden layers contain 64, 128, and 64 neurons respectively, use the Tanh activation function, and introduce a batch normalization layer to improve training stability and model generalization ability; A residual connection is introduced after the second hidden layer, which adds the output of the first hidden layer to the output of the second hidden layer and then passes it to the third hidden layer, effectively alleviating the gradient vanishing problem in deep network training; The output layer produces a 5-dimensional perceptual response vector, including perceptual brightness, hue, saturation, chroma a, and chroma b; The Adam optimizer was used to train the model. The initial learning rate was set to 0.001, and a learning rate decay strategy was used to reduce the learning rate to 90% of the previous value every 50 training cycles. The model performance was evaluated using 5-fold cross validation, and overfitting was avoided using the early stopping strategy.

[0029] In specific application scenarios in the printing industry, this personalized visual perception model can be specially optimized for color-blind observers. For example, for observers with red-green color deficiency, the system will train a dedicated model based on their personal color matching experimental data. This model can predict the perceptual difficulties such observers will have in distinguishing red and green hues, and adjust printing parameters accordingly to enhance the contrast of the red and green channels, allowing color-blind observers to better distinguish key color information in the image. In the printing of a product catalog containing red and green color combinations, the system can automatically identify color combinations that may cause confusion for color-blind observers and make parameter compensation so that the final printed product still maintains good readability and color distinction for special observers.

[0030] Through these steps, a personalized visual perception model is developed that accurately predicts the color perception characteristics of a specific observer under different environmental conditions, providing a personalized data foundation for subsequent parameter optimization. The model's innovation lies in its integration of neural coding theory and color appearance models, adapting to individual differences among observers. This provides technical support for optimizing print quality for special populations, such as those with color deficiency.

[0031] Step 4: Based on the multidimensional color characteristic model, environmental parameter dataset, and personalized visual perception model, an iterative color separation algorithm is used to obtain the optimal ink distribution. The color conversion parameters are adjusted in combination with real-time environmental conditions. The color compensation parameters are fine-tuned through the personalized visual perception model to dynamically adjust the printing parameters.

[0032] Based on the multidimensional color characteristic model, environmental parameter data set, and individual color perception parameters obtained in the previous steps, an iterative color separation algorithm is used to calculate and dynamically adjust the optimal printing parameters to generate an environmentally adaptive printing parameter configuration. This step includes: The optimal ink volume distribution is calculated using an iterative color separation algorithm. , solve the optimal ink volume combination through the following optimization problem : ; in, is the amount of cyan ink, is the amount of magenta ink, For yellow ink, is the amount of black ink, represents the color difference function, represents the established color response model, Indicates finding the minimum value of variables c, m, y and k. Indicates the target color.

[0033] ; in, is the learning rate, is the gradient of the color difference function, represents the ink volume combination of the t+1th iteration, Represents the ink volume combination of the tth iteration. The specific implementation steps of the iterative color separation algorithm are as follows: Initialize ink volume combination , which can be obtained by using an approximation based on a color lookup table or by inverse transformation of the color response model; Calculate the predicted output color under the current ink volume combination: ; in, Indicates that at the tth iteration, the color response model The predicted output color.

[0034] Compute the color difference between the predicted color and the target color: ; in, is the brightness value, and are the first and second chromaticity coordinates in the CIELAB color space, Represents the color difference between the predicted color and the target color at the tth iteration, It represents the square of the difference between the predicted brightness value and the target brightness value. It represents the square of the difference between the predicted a value and the target a value. It represents the square of the difference between the predicted b value and the target b value.

[0035] If the color difference is less than a preset threshold (usually 0.5ΔE units) or the maximum number of iterations (usually 100) is reached, the iteration is stopped; otherwise, continue to step 5; Calculate the gradient of the color difference function with respect to the ink volume combination: ; in, represents the gradient operator, represents the partial derivative, Represent the partial derivatives of variables c, m, y, and K respectively, represents the color difference function; Adopt adaptive learning rate strategy to update ink volume combination, learning rate Dynamic adjustment according to gradient size: ; in, is the learning rate of the tth iteration, is the initial learning rate (usually 0.05), is the attenuation coefficient (usually 0.01), is the current iteration number; Apply an ink level constraint to ensure that the updated ink level value is within the valid range: ; in, and Represent the maximum and minimum functions, respectively. represents the ink volume combination of the t+1th iteration; Return to step 2 and continue iterating.

[0036] In the application scenario of high-quality artwork reproduction and printing, the algorithm can accurately calculate the optimal ink ratio for complex colors.

[0037] Dynamically adjust color conversion parameters based on real-time environmental conditions. Calculate environmental compensation factors based on the environmental parameter data set collected in step 2. : ; in, is the environmental compensation function, is the ambient spectral distribution function, is the ambient illumination, is the ambient color temperature, is the spectral radiance distribution at spatial position s, is the change of the environmental parameter over time t, and the environmental parameter is mapped to the color conversion adjustment coefficient. The adjusted color conversion parameter is: ; in, is the basic color conversion parameter, are the adjusted parameters.

[0038] Environmental compensation function The specific implementation steps are as follows: Build environment parameter vector: ; in, is the average value of the ambient spectrum, is the ambient illumination, is the color temperature, is the coefficient of variation of spatial illumination distribution, is the time rate of change of color temperature; Standardization is used to make each environmental parameter in the same value range: ; in, is the preset mean value, represents the standardized environmental parameter vector, represents the original environment parameter vector, It is represented as an activation function; Calculate the reference vector under reference environmental conditions (usually D65 standard light source, illuminance 500lux) ; Applying a color constancy compensation matrix Calculate the preliminary compensation factor: ; in, represents the preliminary compensation factor, represents the color constancy compensation matrix, represents the reference vector under reference environmental conditions; Apply the von-Kries transform based on the color temperature difference to generate the color adaptation coefficient : ; in, 、 、 are the adaptation coefficients of long-wave, medium-wave, and short-wave cones, respectively, 、 、 They are the response values ​​of the first, second and third cone cells under D65 light source, 、 、 Current color temperature The first, second, and third cone response values ​​under ; Considering the uneven distribution of spatial illumination, calculate the spatial compensation factor: ; in, is the iteration weight coefficient (usually 0.15), is the spatial compensation factor, is the coefficient of variation of spatial illumination distribution; Combining the above factors, the final environmental compensation factor is generated: ; in, is the environmental compensation factor, is the initial compensation factor, Indicates is a diagonal matrix with diagonal elements, is the space compensation factor.

[0039] In retail display applications, this environmental compensation function can adapt to varying shopping mall lighting environments. For example, if the same product catalog is displayed in a fluorescent-lit mall and a boutique illuminated by warm LED lighting, traditional printed materials may exhibit noticeable color deviations under different lighting conditions. Printing parameters optimized using this environmental compensation function can predict and compensate for these lighting variations, ensuring that product images maintain color consistency with the actual product regardless of the commercial environment, improving customer purchasing decisions.

[0040] Fine-tune the color compensation parameters based on the individual perception model to ensure perceptual consistency. Based on the personalized visual perception model constructed in step 3 , calculate the individual compensation factor : ; in, is the individualized compensation function, is an individualized visual perception model. For the target color, is the environmental compensation factor, The individual perception model, target color and environmental compensation factor are mapped into individual adjustment coefficients for individual compensation factors.

[0041] The final optimized printing parameters are: ; in, To adjust printing parameters, is the adjusted parameter, is the individual compensation factor.

[0042] Individualized compensation function The specific implementation steps are as follows: Using environmental compensation parameters , the adjusted parameters calculate the expected output color ,in, To predict color, is the color response function; Applying personalized visual perception models , predict the target observer's perceptual response to the expected output color under the current environmental conditions: ; in, is the predicted perceptual response, is an individualized visual perception model. To predict color, For environmental conditions; Compute the target observer's expected perceptual response to the target color: ; in, For target perception response, is an individualized visual perception model. For the target color, is the reference environmental condition; Compute the perceptual difference vector: ; in, is the perceptual difference vector, For target perception response, for the predicted perceptual response; Designing a mapping matrix from perceptual responses to parameter adjustments ,The matrix is ​​optimized through a large amount of color matching experimental data; Calculate the parameter adjustment coefficient: ; in, Indicates the parameter adjustment amount, represents the mapping matrix from perceptual response to parameter adjustment, represents the perceptual difference vector; Generate individualized compensation factors: ; in, is the individual compensation factor, To adjust the intensity parameter, it can be adjusted according to the accuracy requirements of the actual application, usually between 0.2 and 0.8. is the parameter adjustment amount.

[0043] This personalized compensation function is crucial in medical image printing applications. For example, when a radiologist reviews a printed version of an X-ray or MRI image, subtle grayscale differences within the image may contain crucial diagnostic information. By building a personalized visual perception model for each physician and applying a personalized compensation function, the system optimizes the image's contrast and grayscale distribution based on their unique visual characteristics, enhancing the details they can discern, thereby improving diagnostic accuracy. This is particularly important in situations where digital devices are inconvenient, such as consultations outside the operating room, to ensure that printed medical images accurately convey the same diagnostic information as their digital counterparts.

[0044] Parameter settings are continuously optimized through a closed-loop feedback mechanism. A parameter optimization iteration system is configured, which includes four core modules: parameter generation, parameter application, result evaluation, and parameter correction. Parameters are continuously optimized through the following iterative formula: ; in, is the iteration weight coefficient, is the ideal parameter, For the The actual parameters of the iteration, For the The optimization parameters for the iterations, For the The optimization parameters for the iterations, is the number of iterations.

[0045] Through the above steps, we achieve optimization and dynamic adjustment of printing parameters based on a multidimensional color characteristic model, real-time environmental parameters, and an individualized perception model. The innovation of this process lies in the integration of physical color models, environmental factors, and individual perception characteristics. Through a closed-loop feedback mechanism, we achieve continuous optimization of parameters, ensuring that the printed results appear consistent to different observers under different environmental conditions.

[0046] Step 5: Based on the dynamic adjustment of printing parameters, the printing output is monitored in real time during the printing process and the printing parameters are corrected to minimize the perceived printing differences.

[0047] During the printing process, a spectral detection sensor system is deployed to monitor and correct the printed output in real time, generating corrected printing parameters. This step includes: The spectral detection sensor system is configured to monitor the printed output in real time. The system includes a high-precision spectral scanning unit, an image analysis unit, and a correction calculation unit, and can collect color output spectral data during the printing process at a frequency of more than 10 times per second. ; Calculate the difference between the actual output and the expected perception. Based on the individualized visual perception model established in step 3, calculate the perceptual representation of the actual output under given environmental conditions. , and with the expected perceptual representation Compare and calculate perceived differences : ; in, is the perceptual difference vector, is the distance function in the perceptual space, For actual perceptual representation, To predict perceptual representation, weighted Euclidean distance is usually used: ; in, is the dimension of perceptual features, is the weight of each dimension, and are the first and second perceptual representation vectors to be compared, and are vectors and No. A quantity, is the total number of components; Apply real-time correction algorithms to automatically adjust printing parameters to minimize perceived differences. Based on calculated perceived differences , adjust printing parameters , generate the corrected parameters : ; in, is the corrected parameter, To optimize the parameters, is the correction factor, To correct the mapping function, the perceptual difference is mapped to the parameter adjustment amount, is the perceptual difference vector. The correction mapping function is implemented by the following neural network model: ; in, is the activation function, 、 are the third and fourth weight matrices respectively, 、 are the fifth and sixth bias vectors respectively, is the correction mapping function, is the perceptual difference vector; Establish an optimized data recording system to collect information such as parameters, environmental conditions, correction operations and final quality evaluation during the printing process to form a structured database The database is used for subsequent system performance analysis and optimization algorithm improvement. The specific data items recorded include: ; in, To optimize the data record collection, For the Initial optimization parameters for the printing task, For environmental conditions, To perceive the difference, is the corrected parameter, For the final quality score, is the total number of printing tasks recorded, The task number.

[0048] Through these steps, a complete print output verification and correction system is implemented. This system monitors color output in real time during the printing process, calculates perceptual differences based on individualized perceptual models, and dynamically adjusts printing parameters using intelligent correction algorithms to minimize these differences. This ensures that the final print presents the intended visual effect for the target observer under specific circumstances. Furthermore, an optimized data recording system is established to provide data support for the system's continuous improvement.

[0049] This embodiment achieves the following technical effects through the implementation of the above five core steps: Improved color reproduction accuracy: By building a high-precision multidimensional color characteristic model, combined with real-time spectral data monitoring and correction, this technology reduces color differences, improves color reproduction accuracy, and enhances color consistency. Compared to traditional color management technologies, this solution can more accurately predict and control color output under different printing conditions, significantly improving color reproduction accuracy.

[0050] Environmental Adaptation: By collecting real-time environmental perception data and dynamically adjusting parameters based on this data, the color perception of printed products remains consistent under varying lighting conditions. Tests have shown that across a range of color temperatures from 2000K to 10000K, this solution produces a perceptual change of no more than 5% in printed products, while traditional methods show a modest improvement under the same conditions. This effectively addresses the limitations of traditional fixed calibration methods in varying environments.

[0051] Meeting Individual Needs: By establishing a personalized visual perception model, we provide personalized printing parameter optimization for observers with different color perception characteristics. Tests have shown that personalized optimization for color-blind observers significantly improves color recognition accuracy, and for average observers, personalized optimization also leads to a certain increase in subjective satisfaction. This ensures that each observer receives the expected visual experience that matches their individual characteristics.

[0052] Enhanced printing stability: Through real-time monitoring and dynamic correction systems, stable color output quality is maintained despite fluctuations in printing conditions (such as temperature, humidity, and paper characteristics). Experiments have shown that this solution significantly improves color output stability under fluctuating environmental conditions, significantly reducing fluctuations in print quality caused by these conditions.

[0053] Improved production efficiency: This solution reduces rework and adjustment time caused by color issues, reducing material waste and improving overall production efficiency. Actual application data shows that the printing scrap rate has been reduced, color adjustment time has been reduced by 50%, and overall production efficiency has been significantly improved.

[0054] Achieving consistent perceptual experience: This solution extends printing parameter optimization from pursuing consistency in physical indicators to perceptual consistency, achieving a fundamental shift in print quality evaluation. This ensures that printed products present optimal visual effects in a variety of practical usage environments, enhancing user experience and product value.

[0055] In summary, this implementation achieves an all-round improvement in printing quality through the synergy of five key steps: multi-dimensional color characteristic modeling, environmental perception data collection, individualized visual perception modeling, parameter optimization and dynamic adjustment, and output verification and correction. It not only addresses the limitations of traditional color management technology, but also creates a new paradigm for printing quality evaluation centered on perceptual experience.

[0056] Application examples of this embodiment This implementation has been put to practical use in a high-end art catalog printing project. This project required accurate color reproduction of artworks under various lighting conditions (including natural lighting in the art gallery, professional LED display lighting, and general home lighting), while also meeting the color perception needs of different viewers (including art professionals and the general public). The following details the specific application process and actual results of this technical solution in this project.

[0057] Implementation example of multidimensional color feature modeling In this project, a spectrometer from a certain company was used to measure color spectrum data under different printing conditions. The specific test conditions and parameter settings are shown in Table 1.

[0058] Table 1: Spectral measurement conditions and parameter settings

[0059] A piecewise polynomial interpolation algorithm was applied to the collected spectral data, dividing the wavelength range into five subintervals. Each subinterval was modeled using a fourth-order polynomial. The model was trained using a stochastic gradient descent algorithm with a learning rate of 0.01 and 500 iterations. Table 2 shows a comparison of model accuracy under different printing conditions.

[0060] Table 2: Comparison of color characteristic model accuracy under different printing conditions

[0061] As can be seen from Table 2, compared with the traditional color model, the multidimensional color characteristic model has achieved an accuracy improvement of more than 60% under various printing conditions. The improvement effect is particularly obvious under complex paper and ink combinations and environmental conditions.

[0062] An example of environmental perception data collection: To address the diverse environments in which artworks are displayed, this project employed a distributed perception network consisting of 12 miniature spectral sensors to monitor lighting conditions in different display areas in real time. Table 3 shows the configuration parameters of the sensor array.

[0063] Table 3: Spectral sensor array configuration parameters

[0064] Using this sensor array, the system can capture changing lighting conditions across different areas of the museum. Based on real-time collected environmental parameters, the system constructs a model of environmental parameter changes, predicting short-term lighting trends and providing this information to the parameter optimization module as a basis for decision-making. Experiments have shown that this model can predict lighting trends within 10 minutes with over 95% accuracy.

[0065] An example of personalized visual perception modeling: In the art catalog printing project, personalized visual perception models were constructed for two key user groups: art professionals and general audiences. Table 4 shows the design parameters for the color matching experiment.

[0066] Table 4: Color matching experiment design parameters

[0067] Based on the collected color matching data, a neural encoding model and a personalized visual perception model were constructed. The performance comparison of the two models in predicting the color perception characteristics of the subjects is shown in Table 5.

[0068] Table 5: Performance comparison of individualized visual perception models

[0069] The personalized visual perception model has demonstrated remarkable effectiveness in specific audience groups. For example, for an art critic with red-green color deficiency, the system trained a dedicated model using their personal color matching data. When processing a reproduction of the famous painting "Sunflowers," which contains subtle differences in red and green tones, the system optimized printing parameters, significantly improving the accuracy of key tones, enabling the critic to more accurately discern and evaluate subtle color variations in the work.

[0070] To illustrate the practical application of personalized visual perception modeling, we'll use a specific case study below. In a high-end art album printing project, we implemented personalized visual perception modeling across the entire process for two key user groups: professional art critics and casual art enthusiasts.

[0071] First, we designed a comprehensive color matching experiment. This experiment used a standard color sample set consisting of 60 representative colors commonly found in works of art. These color samples covered a wide range, from highly saturated primary colors to subtle neutral tones. The experiments were conducted under three different lighting conditions: simulated natural museum lighting (5500K), professional LED display lighting (3800K), and ambient home lighting (2700K).

[0072] The subjects in the experiment included Mr. P, a color-blind art critic with 20 years of experience. Mr. P reviews Impressionist artworks, but his red-green color deficiency consistently affects his ability to accurately perceive certain hue variations. Preliminary testing revealed that, under normal lighting conditions, Mr. P's ability to discern color differences on the red-green axis was approximately 40% lower than that of observers with normal color vision, but his ability to discern color differences on the yellow-blue axis was comparable to that of normal observers.

[0073] To build Mr. P’s personalized visual perception model, we collected the following data: Standard color vision test results to determine the type and degree of Mr. P's color deficiency; Color matching results for 60 color samples under three lighting conditions, with each sample tested five times; Data on the effects of different background colors on Mr. P’s color perception; Data on changes in Mr. P's visual adaptation state after different observation times.

[0074] Based on these data, we used a three-stage modeling approach to construct Mr. P’s personalized visual perception model: Phase 1: Basic cone response model, using the modified LMS response function, describes Mr. P's cone cell response characteristics through the following equation: ; ; .

[0075] in, 、 and They represent Mr. P’s long-wavelength, medium-wavelength, and short-wavelength cone cell responses, respectively. 、 and These correction coefficients are derived by reverse deduction from color matching experimental data.

[0076] The second stage: neural coding layer modeling, using a multi-layer perceptron network with environmental adaptability. The network structure includes: Input layer: 7 neurons, receiving cone response values ​​and environmental parameters; First hidden layer: 32 neurons, using ReLU activation function; Second hidden layer: 64 neurons, using Sigmoid activation function; Output layer: 5 neurons corresponding to perceived brightness, hue, saturation, and chromaticity coordinates.

[0077] The network is trained by minimizing the perceptual prediction error using a stochastic gradient descent algorithm with a batch size of 32 and an initial learning rate of 0.01, and converges after 2000 epochs.

[0078] The third phase, perceptual calibration and validation, evaluated model performance using a set-aside test dataset and calibrated the model through subjective evaluation of actual printed samples. During the calibration process, Mr. P rated a series of printed samples under different conditions. The system recorded the differences between the scoring results and the model's predictions and fine-tuned the model parameters accordingly. Table 6 shows a comparison of the color prediction accuracy of Mr. P's personalized visual perception model with a traditional model.

[0079] Table 6: Performance evaluation of Mr. P’s personalized visual perception model

[0080] An example of parameter optimization implementation: In an artwork catalog printing project, an iterative color separation algorithm was applied based on the previously established multidimensional color characteristic model, environmental perception data, and personalized visual perception model to achieve dynamic optimization of printing parameters. Table 7 lists the key indicators of the parameter optimization process.

[0081] Table 7: Key indicators of printing parameter optimization process

[0082] The algorithm’s optimization strategies for different printing areas vary. Table 8 shows the specific optimization parameters for three typical image areas.

[0083] Table 8: Parameter optimization strategies for different image regions

[0084] In practice, the system dynamically adjusts printing parameters based on the lighting conditions of different exhibition environments. For example, if it detects a change in exhibition hall lighting from natural light (5500K) to LED lighting (3200K), the system automatically recalculates the optimal printing parameters within 2.4 seconds, adjusting the CMYK values ​​to ensure accurate color reproduction in the new environment, with color difference within a range of ΔE2000 < 2.

[0085] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make more forms of equivalent embodiments based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. A method for improving printing quality based on printing parameter optimization, characterized in that: The following steps are involved: Build a multidimensional color characteristic model using spectral data. The multidimensional color characteristic model includes training data on printing conditions, equipment characteristics, ink characteristics, and paper characteristics. Based on the construction of a multi-dimensional color characteristic model, the lighting conditions and related parameters of the observed environment are collected in real time through spectral sensors to generate an environmental parameter data set; Based on the environmental parameter dataset, a personalized visual perception model of the observer is established through color matching experiments. The personalized visual perception model maps physical stimuli to perceptual responses. Based on a multidimensional color characteristic model, an environmental parameter dataset, and a personalized visual perception model, an iterative color separation algorithm is used to obtain the optimal ink distribution. Color conversion parameters are adjusted in accordance with real-time environmental conditions, and color compensation parameters are fine-tuned using a personalized visual perception model to dynamically adjust printing parameters. Based on dynamic adjustment of printing parameters, the printing output is monitored in real time during the printing process and the printing parameters are corrected to minimize the perceived printing differences.

2. The method for improving printing quality based on printing parameter optimization according to claim 1, characterized in that: The step of constructing a multidimensional color characteristic model based on spectral data includes: collecting color output spectrum data of a printing device under different printing conditions, wherein the printing conditions include temperature, humidity, paper type, and ink type; Use piecewise polynomial interpolation algorithm to model spectral data and construct color space mapping relationship; A comprehensive color response model including device characteristics, ink characteristics and paper characteristics is constructed, wherein the comprehensive color response model is expressed as a product of the device response function, the ink response function and the paper response function.

3. The method for improving printing quality based on printing parameter optimization according to claim 2, characterized in that: The piecewise polynomial interpolation algorithm divides the wavelength range into multiple subintervals, applies a different polynomial function to each subinterval, and obtains the polynomial coefficients by solving the least square method.

4. The method for improving printing quality based on printing parameter optimization according to claim 1, characterized in that: The step of using a spectral sensor to collect lighting conditions and related parameters of the observation environment in real time includes: Configure a micro-spectral sensor array system to measure real-time environmental spectral parameters; Collect environmental spectral distribution data, measure key parameters of ambient illumination and color temperature, and obtain chromaticity coordinates and color purity; Use distributed sensor arrays to capture spatial illumination unevenness and generate a spatial illumination distribution model; The collected environmental data is analyzed in time series to generate an environmental parameter change model to predict the short-term change trend of the ambient lighting conditions.

5. The method for improving printing quality based on printing parameter optimization according to claim 4, characterized in that: The micro-spectral sensor array system includes multiple distributed spectral sensor units and a central data processing unit. Each sensor unit captures spectral information within a wavelength range of 380nm to 780nm and transmits data to the central processing unit.

6. The method for improving printing quality based on printing parameter optimization according to claim 1, characterized in that: The step of establishing the observer's personalized visual perception model through the color matching experiment includes: Perform color matching experiments to obtain observer-specific color perception characteristics data; Apply neural coding theory to establish a mathematical mapping relationship from physical stimuli to perceptual responses; Combined with color appearance models to obtain color perception properties under specific conditions; Individual perception data is modeled through machine learning algorithms to train personalized visual perception models.

7. The method for improving printing quality based on printing parameter optimization according to claim 6, characterized in that: The personalized visual perception model adopts a multi-layer perceptron structure, including: An input layer that receives physical stimulus parameters and environmental conditions; Three hidden layers, containing 64, 128, and 64 neurons respectively, using a nonlinear activation function; An output layer that produces predicted individual perceptual responses, including perceived brightness, hue, saturation, and chromaticity values.

8. The method for improving printing quality based on printing parameter optimization according to claim 1, characterized in that: The dynamic adjustment of printing parameters includes: The iterative color separation algorithm is used to solve the optimization problem of the optimal ink volume combination; Obtain environmental compensation factors based on real-time environmental conditions and adjust color conversion parameters; Combined with the individualized visual perception model, the individualized compensation factor is obtained and the color compensation parameters are fine-tuned; Parameter settings are continuously optimized through a closed-loop feedback mechanism.

9. The method for improving printing quality based on printing parameter optimization according to claim 1, characterized in that: The steps of real-time monitoring of printing output and correcting printing parameters include: Configure the spectrum detection sensor system to monitor the printing output in real time; Obtain the difference between actual output and expected perception based on an individualized visual perception model; Apply real-time correction algorithms to automatically adjust printing parameters to minimize perceived differences; Establish an optimized data recording system to collect information on parameters, environmental conditions, correction operations and final quality evaluation during the printing process.

10. A computer-readable storage medium, characterized in that Used to store computer-readable instructions, when the computer-readable instructions are read by a computer, a printing quality improvement method based on printing parameter optimization as described in any one of claims 1 to 9 can be executed.

Images