Dynamic digital brush generation method and system for realizing Chinese painting ink shading effect

Abstract

The invention discloses a dynamic digital brush generation method and system for realizing a Chinese painting ink shading effect, and relates to the technical field of Chinese painting ink shading. According to the method, a multi-dimensional blooming effect generation model is constructed, physical simulation and technical rules are fused, complex and changeable Chinese painting ink blooming situations under different paper and humidity conditions are accurately simulated, and the blooming effect in digital painting is closer to the charm of a traditional Chinese painting; a mapping relation between user operation parameters and shading effect parameters is established, and training is performed by utilizing a machine learning algorithm, so that the system can make sensitive and fine feedback on the operation of the user, and rich changes caused by light and slow strokes in Chinese painting creation are reflected; the layer-by-layer permeation process of ink color between Chinese art paper fibers is simulated, smooth transition is carried out on color values of adjacent pixels, and a stiff boundary caused by traditional linear interpolation is avoided.

Description

Technical Field

[0001] This invention relates to the field of traditional Chinese ink wash painting technology, specifically to a method and system for generating dynamic digital brushes to achieve the ink wash effect in traditional Chinese painting. Background Technology

[0002] Traditional Chinese painting (also known as Chinese scroll painting) is a traditional Chinese painting form, primarily referring to scroll paintings painted on silk, Xuan paper, or other similar materials and then mounted. Ink wash technique is an important aspect of traditional Chinese painting, its core being the use of the penetrating and diffusing properties of ink to achieve color gradation and create atmosphere. Washing refers to the natural spreading of ink or color on Xuan paper, using a wet-on-wet technique to create gradual transitions in color intensity, forming a visual effect of light and shadow. This technique emphasizes the soft transition of colors and a sense of spatial depth, and is commonly found in landscapes, flowers, and birds.

[0003] Dynamic digital brushes are a type of brush in digital painting tools that features adjustable parameters and variable behavior. They generate natural and flexible brushstroke effects by responding in real time to user actions (such as pressure sensitivity, speed, and angle) or external settings (such as texture and blending modes). Compared to traditional static brushes, their core value lies in simulating the texture of realistic painting (such as the lightness and heaviness of a brush or the diffusion of watercolor) and improving creative efficiency (such as reducing repetitive operations through dynamic parameters).

[0004] Existing methods and systems for generating dynamic digital brushes to achieve ink wash effects in traditional Chinese painting have several shortcomings. Some methods can only simulate relatively simple wash patterns, failing to accurately represent complex and varied ink wash scenarios, such as the effects under different paper types and humidity conditions. Furthermore, in terms of dynamic response, existing systems lack sensitivity and subtlety in their feedback to user operations, failing to adequately reflect the rich variations in brushstroke pressure and speed in traditional Chinese painting. Additionally, the naturalness of color transitions is often lacking; many systems produce uneven color gradients that fall short of the soft, natural color transitions characteristic of traditional Chinese painting. Adjusting brush parameters is often inconvenient and unintuitive, requiring creators to spend considerable time figuring things out and fine-tuning, which impacts the smoothness and efficiency of the creative process. Moreover, existing systems suffer from compatibility issues with different digital painting software, limiting their application scope. Therefore, we propose a method and system for generating dynamic digital brushes to achieve ink wash effects in traditional Chinese painting. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for generating dynamic digital brushes to achieve the ink wash effect in traditional Chinese painting, in order to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention specifically adopts the following technical solution:

[0007] A method for generating dynamic digital brushes to achieve the ink wash effect in traditional Chinese painting includes the following steps:

[0008] S1. Construct a multi-dimensional model for generating ink diffusion effects: Collect the fiber structure parameters of traditional Xuan paper and the physical properties of ink, establish a simulation model of ink diffusion in Xuan paper, extract the rules of traditional ink diffusion techniques, integrate physical simulation with the rules of techniques, and establish a multi-dimensional model for generating ink diffusion effects.

[0009] S2. Dynamic Parameter and Blur Effect Mapping: Establish a mapping relationship between user operation parameters and blur effect parameters, train the mapping model through machine learning algorithms, and realize the real-time dynamic response of user operation to blur effect;

[0010] S3. Design a real-time rendering engine: Based on GPU acceleration technology, it performs real-time rendering and integrates ink texture synthesis, paper texture mapping, and color mixing algorithms. During the real-time rendering process, it combines the blurring effect generation model and dynamic parameter mapping results. According to different user operations and preset blurring effect requirements, it calculates the color value and texture information of each pixel to realize the real-time generation and display of the blurring effect.

[0011] S4. Parameter Optimization and Compatibility Adaptation: Iteratively optimize parameter mapping relationships through user feedback to improve the sensitivity and finesse of operation response; establish a universal interface to achieve compatibility adaptation between the system and different software.

[0012] S5. Effect Evaluation and Continuous Iteration: Construct a multi-dimensional blending effect evaluation system, combine subjective evaluation and objective indicators to verify the authenticity and artistry of the generated effect, and continuously iterate and optimize the model and parameters based on the evaluation results to improve the system's generation quality and user experience.

[0013] Furthermore, the fiber structure parameters include fiber density and porosity; the physical properties of the ink include ink concentration and permeability coefficient; and the rules of ink blending techniques include multi-layer superposition in the ink accumulation method and dry-wet alternation in the ink breaking method.

[0014] Furthermore, the operating parameters include pressure sensitivity, pen stroke speed, and pen tip angle; the blurring effect parameters include diffusion range, concentration gradient, and edge blurring degree; the machine learning algorithm is either a support vector machine or a neural network.

[0015] Furthermore, the ink texture includes ink stain texture and smudged edge texture; the paper texture includes the smooth texture of sized Xuan paper and the rough texture of unsized Xuan paper; the color mixing algorithm uses a physics-based layered overlay algorithm or a nonlinear color interpolation algorithm. The layered overlay algorithm simulates the process of ink penetrating layer by layer between Xuan paper fibers, dividing the ink of each stroke into a base layer, a transition layer, and a surface layer, and adjusting the concentration and transparency of each layer to achieve a natural sense of layering from dark to light. The nonlinear color interpolation algorithm addresses the color gradation requirements of smudged edges by dynamically adjusting the curvature of the interpolation function to smoothly transition the color values ​​of adjacent pixels, avoiding the harsh boundaries caused by traditional linear interpolation and enhancing the softness and realism of the smudged edges.

[0016] Furthermore, the general interface uses a plug-in interface that supports mainstream digital painting software such as Photoshop and Procreate.

[0017] Furthermore, the subjective evaluation includes inviting professional Chinese painters to score the technique reproduction and artistic expression of the blending effect on a ten-point scale, and collecting feedback from ordinary users on the operation response speed, effect controllability, and interface friendliness through online questionnaires; the objective indicators include the smoothness of the blending edge, the richness of the ink color layering, and the frame rate stability of real-time rendering.

[0018] A dynamic digital brush generation system for achieving the ink wash effect in traditional Chinese painting includes:

[0019] Ink wash model building unit: used to collect the physical parameters of Xuan paper and the property parameters of ink, integrate traditional techniques and rules, and build a physical-technical fusion model for generating wash effects;

[0020] Dynamic parameter mapping unit: used to establish the mapping relationship between user operation parameters and blurring effect parameters, and realize real-time dynamic response through machine learning model;

[0021] Real-time rendering unit: Used for real-time rendering of ink wash texture synthesis, paper texture mapping, and color mixing based on GPU acceleration, generating natural blurring effects;

[0022] Interactive interface unit: Based on a touch screen, it provides an intuitive parameter adjustment interface, allowing users to quickly set the blurring effect parameters and preview the effect in real time;

[0023] Compatibility adapter unit: Used to achieve system compatibility with different digital painting software through a universal interface, thereby expanding the application scope of the system.

[0024] Furthermore, the ink wash model construction unit includes:

[0025] Xuan paper physical parameter acquisition submodule: By analyzing the microstructural characteristics of Xuan paper fiber density and interlacing morphology through high-resolution microscopic images, and combined with ink absorption experiments under constant temperature and humidity environment, parameters such as ink absorption rate, diffusion coefficient and water evaporation rate are measured to establish a digital physical model of Xuan paper material, so as to achieve accurate simulation of ink absorption characteristics of different types of Xuan paper.

[0026] The ink property parameter acquisition submodule uses a spectrometer to obtain the spectral reflectance of ink at different dilution ratios, converts it into quantized values ​​in RGB and HSV color spaces, measures the dynamic viscosity of ink using a rotational viscometer, and obtains the diffusion coefficient of ink in liquid by combining diffusion experiments. It constructs a correlation model between ink concentration, viscosity and diffusion characteristics, and supports accurate reproduction of five ink color levels: "burnt, thick, heavy, light and clear".

[0027] The traditional techniques and rules integration submodule extracts the rule parameters of the core techniques of ink accumulation, ink breaking, and ink blending through motion capture and interviews with traditional Chinese painters. These parameters are then transformed into algorithmic logic based on probability statistics or machine learning. This logic is then used in real-time collaborative calculations with the physical model of Xuan paper and the ink property model to form a dual-driven ink blending effect generation mechanism that combines physical properties and technique rules.

[0028] Furthermore, the dynamic parameter mapping unit includes:

[0029] The user operation parameter perception submodule uses a digital tablet pressure sensor, an intelligent pen posture capture device, and a user interface to collect dynamic operation parameters in real time, such as the peak pressure, pen speed change rate, pen tip tilt angle, pen stroke continuity, and dwell time. At the same time, it acquires static parameters preset by the user, such as the type of Xuan paper, ink concentration, and blending method. It normalizes multi-source heterogeneous data into a unified format operation parameter matrix to achieve accurate quantification of user intent.

[0030] The sub-module for correlation of ink diffusion effect parameters: Based on the effect parameters such as diffusion radius, ink gradient distribution coefficient, blurring boundary ambiguity, and ink layer thickness generated by the ink diffusion model, the module mines the potential correlation between user operation parameters and ink diffusion effect parameters through mutual information entropy calculation or nonlinear principal component analysis, establishes an operation and effect correlation feature library, and uses a support vector machine classifier to label feature weights to distinguish the degree of influence of core operation parameters and secondary operation parameters on the effect.

[0031] The real-time mapping model construction submodule uses a lightweight Transformer model combined with a deep reinforcement learning framework to build a real-time mapping model. It takes the user operation parameter matrix as input and the shading effect association features as output. The input sample data is pre-trained, and the model parameters are updated in real time through an online learning mechanism to achieve the effect parameter response within 50 milliseconds when the user operation changes. This ensures that the shading effect is presented naturally with the dynamic adjustment of the user's pen movements, achieving an interactive experience where the pen moves with the heart.

[0032] Furthermore, the real-time rendering unit includes:

[0033] The ink wash texture synthesis submodule utilizes a GPU-based CUDA / Metal parallel computing framework. It combines parameters such as the diffusion radius and ink gradient distribution coefficient output from the ink wash model with conditions for generative adversarial networks to pre-train an ink wash texture generation model. This generates ink wash textures that match the current diffusion state in real time. The model uses diffusion effect parameters as input and generates pixel-level ink density distributions through GPU-accelerated depth convolution operations. Worley noise is introduced to simulate the random resistance effect of Xuan paper fibers on ink, enhancing the natural randomness of the texture and ensuring a rendering frame rate of over 30fps at 4K resolution, meeting the real-time requirements of professional creation.

[0034] Paper Texture Mapping Submodule: Utilizing the digital model of fiber density and interlacing morphology constructed by the Xuan paper physical parameter acquisition submodule, the geometric texture of Xuan paper fibers is generated through 3D point cloud reconstruction technology and converted into normal maps and displacement maps; through the GPU texture mapping pipeline, the paper texture and ink texture are fused in multiple channels to simulate the difference in fiber adsorption of ink color, and at the same time, the diffuse reflection and specular reflection characteristics of the paper surface are adjusted through a micro-surface-based BRDF model to restore the texture of Xuan paper;

[0035] The color mixing calculation submodule, based on traditional techniques and ink properties, uses the Beer-Lambert light absorption law combined with Kubelka-Munk theory to construct a physical color mixing model, which calculates the superposition effect of multiple ink layers in real time. The model converts each layer of ink into the Lab color space and performs non-linear mixing according to the technique of dry ink base and wet ink glazing. At the same time, it dynamically adjusts the mixing weight by combining the ink absorption rate of Xuan paper and the water evaporation parameters, and realizes pixel-by-pixel color calculation through the fragment shader of the GPU to ensure that the transition of ink from dark to light is natural and smooth, avoiding color block breaks or color overflow, and truly restoring the layered effect of ink with five colors in traditional ink wash.

[0036] The beneficial effects of this invention are as follows:

[0037] 1. This invention constructs a multi-dimensional model for generating ink wash effects, integrating physical simulation with technical rules to accurately simulate the complex and varied ink wash scenarios in traditional Chinese painting under different paper types and humidity conditions, making the ink wash effects in digital painting closer to the charm of traditional Chinese painting.

[0038] 2. This invention establishes a mapping relationship between user operation parameters and blending effect parameters, and uses machine learning algorithms for training, enabling the system to provide sensitive and delicate feedback to user operations, and well reflect the rich variations brought about by the lightness and heaviness of brushstrokes in traditional Chinese painting. When users adjust pressure intensity, brush speed, brush tip angle, etc. during brush strokes, the system can adjust the blending effect in real time in a short period of time, realizing a creative experience where the brush moves with the heart.

[0039] 3. This invention uses a physics-based layered overlay algorithm or a nonlinear color interpolation algorithm to perform color mixing calculations, simulating the process of ink penetrating layer by layer between Xuan paper fibers, and smoothly transitioning the color values ​​of adjacent pixels, avoiding the harsh boundaries caused by traditional linear interpolation; the generated shading color gradient is smooth and natural, highly consistent with the soft and natural color transition effect in traditional Chinese painting, enhancing the softness and realism of the shading edges.

[0040] 4. The interactive interface unit based on the touch screen of this invention provides an intuitive parameter adjustment interface, which supports users to quickly set the blurring effect parameters and preview the effect in real time. Creators do not need to spend a lot of time exploring and debugging brush parameters, and can create more smoothly, which greatly improves the creation efficiency.

[0041] 5. This invention uses a multi-dimensional evaluation system for the blending effect, combining subjective evaluation with objective indicators to verify the generated effect. Based on the evaluation results, the model and parameters are continuously iterated and optimized to ensure that the system can continuously improve the quality of generation and user experience, providing creators with a higher quality and more efficient tool for creating traditional Chinese ink wash paintings, and promoting the development of digital painting in the field of traditional Chinese painting. Attached Figure Description

[0042] Figure 1 This is a flowchart of the method of the present invention;

[0043] Figure 2 This is a schematic diagram of the system modules of the present invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0045] Please see Figure 1This invention provides a method for generating dynamic digital brushes to achieve the ink wash effect in traditional Chinese painting, comprising the following steps:

[0046] S1. Construct a multi-dimensional blurring effect generation model: Collect the fiber structure parameters of traditional Xuan paper and the physical properties of ink, establish a diffusion simulation model of ink in Xuan paper, and extract the rules of traditional blurring techniques. Integrate physical simulation with the rules of techniques to establish a multi-dimensional blurring effect generation model. This model can accurately simulate the penetration rate, blurring range, and ink density evolution caused by water evaporation in the gaps between Xuan paper fibers. At the same time, the operational logic of traditional blurring techniques such as flat dyeing, baking dyeing, and broken dyeing is integrated into the model parameters. This ensures that the blurring effect when drawing with a digital brush not only conforms to the physical laws of the interaction between ink and Xuan paper, but also restores the sense of layering of ink with five colors and the freehand style of broken brushstrokes in traditional techniques. This provides a blurring effect that is both realistic and traditional for the real-time generation of subsequent dynamic digital brushes.

[0047] S2. Dynamic Parameter and Blending Effect Mapping: Establish a mapping relationship between user operation parameters and blending effect parameters. Train the mapping model through machine learning algorithms to achieve real-time dynamic response from user operation to blending effect. Furthermore, based on the pen pressure, speed, angle, and operation details such as pausing and sweeping strokes when the user wields the pen, adjust parameters such as penetration rate, ink diffusion coefficient, and water evaporation rate in the blending effect generation model in real time. This allows the digital brush to present a blending dynamic like a real calligraphy brush during the drawing process, with subtle changes in user operation.

[0048] S3. Real-time rendering engine design: Based on GPU acceleration technology, it performs real-time rendering, integrating ink texture synthesis, paper texture mapping, and color mixing algorithms (such as simulating the layering of ink overlays and the softness of color gradients). During real-time rendering, it combines the blurring effect generation model and dynamic parameter mapping results to calculate the color value and texture information of each pixel according to different user operations and preset blurring effect requirements, realizing the real-time generation and display of blurring effects. This allows users to see the dynamic evolution of ink slowly spreading along the brushstroke trajectory and water penetrating into the gaps between the Xuan paper fibers in real time, just like the natural feedback when a real brush is used to paint on Xuan paper, making the digital creation process more immersive. At the same time, it supports high-resolution canvases and complex brushstroke combinations. Even when drawing large landscape scrolls or delicate flower and bird paintings, it can maintain the smoothness of rendering without lag due to complex details. It provides professional digital Chinese painting creators with an efficient and realistic real-time drawing environment, ensuring that the blurring effect of each stroke can be accurately presented and adjusted in real time.

[0049] S4. Parameter Optimization and Compatibility Adaptation: Parameter mapping relationships are iteratively optimized through user feedback to improve the sensitivity and finesse of operation response; a universal interface is established to achieve compatibility adaptation between the system and different software; furthermore, the mapping curves of pen pressure and ink density, as well as the response thresholds of speed and blur range, can be continuously adjusted based on user feedback, making the operation feel of the digital brush closer to that of a real calligraphy brush. Even veteran artists accustomed to traditional painting can quickly adapt and accurately control the blurring effect; at the same time, the establishment of the universal interface breaks down software barriers, supporting seamless integration with mainstream digital painting software such as Photoshop, Procreate, and Clip Studio Paint. The dynamic digital brush of this invention can be directly invoked without the need for additional plugin installation or setting adjustments, meeting the software usage habits of different creators and expanding the application scenarios of the technology.

[0050] S5. Effect Evaluation and Continuous Iteration: Construct a multi-dimensional evaluation system for ink blending effects, combining subjective evaluation with objective indicators to verify the authenticity and artistry of the generated effects. Based on the evaluation results, continuously iterate and optimize the model and parameters to improve the system's generation quality and user experience. Through this closed-loop mechanism of feedback, evaluation, and iteration, continuously narrow the difference between digital brushes and real brushes in ink blending effects, making the ink blending closer to the artistic essence of traditional Chinese painting. At the same time, continuously adapt to users' higher demands for operation feel and effect precision, ensuring that the system always maintains technological advancement and superior user experience, providing digital Chinese painting creation with a more traditional and vibrant tool support.

[0051] In this embodiment, preferably, the fiber structure parameters include fiber density and porosity. Collecting parameters such as fiber density and porosity allows for accurate reproduction of the physical structure characteristics of real brush fibers, providing underlying structural data support for the dynamic blending model. The physical properties of ink include ink concentration and permeability coefficient. Collecting parameters such as ink concentration and permeability coefficient allows for accurate simulation of the flow, penetration, and diffusion characteristics of real ink on Xuan paper, providing key data input for the physical properties of ink in the dynamic blending model, ensuring accurate digital brush drawing. The blending technique rules include multiple methods of ink accumulation. The layering and wet-drying techniques of the "breaking ink" method are key features. The layering of multiple ink layers in the "accumulating ink" method simulates the absorption and retention of ink by fibers during layering in real painting, achieving a gradual transition from light to dark. Each layer follows the cumulative effect of fiber pores, preventing the bottom layer from being completely covered. Instead, the multiple layers create a thicker, more textured ink layer, avoiding the flattening problem common in digital painting. The wet-drying technique of the "breaking ink" method simulates the interaction between dry brush and wet paper, and wet brush and dry paper, by dynamically adjusting the interaction between brush humidity, ink concentration, and paper moisture. The rules of blending techniques, fiber structure parameters, and the physical properties of ink and water together constitute the core driving force of the dynamic blending model, enabling digital brushes to replicate the feel of operating a real calligraphy brush.

[0052] In this embodiment, preferably, the operation parameters include pressure sensitivity, stroke speed, and brush tip angle. Through real-time acquisition and dynamic mapping of pressure sensitivity, stroke speed, and brush tip angle, the operation force, brush rhythm, and brush tip turning characteristics during real brush painting are accurately simulated. The blurring effect parameters include diffusion range, density gradient, and edge blurring degree. Through real-time acquisition and dynamic adjustment of diffusion range, density gradient, and edge blurring degree, the natural diffusion boundary, the gradual transition of ink color from the center to the edge, and the soft blurring effect of the blurring edge during real ink painting are accurately reproduced. This provides a real-time effect control basis for the dynamic blurring model, ensuring the sense of layering and realism of the blurring area in digital drawing, and avoiding harsh boundaries or uniform color. The design employs a block distribution approach. The machine learning algorithm utilizes either Support Vector Machines (SVMs) or Neural Networks. SVMs, with their superior high-dimensional feature nonlinear mapping capabilities, can accurately establish complex correspondences between operational parameters such as pressure sensitivity and brushstroke speed, and parameters related to the diffusion range and concentration gradient of the ink wash effect. This effectively addresses the subtle operational differences that traditional linear models cannot capture. Neural Networks (such as Convolutional Neural Networks or Long Short-Term Memory Networks), through feature extraction and dynamic learning from multiple layers of neurons, can autonomously learn the dynamic laws of ink penetration and diffusion from a large amount of real-world ink wash process image sequences and brushstroke pressure time-series data. This enables accurate prediction of the blurring degree of the ink wash edge and changes in ink color gradient. By embedding machine learning algorithms into the core computational stage of the dynamic ink wash model, the system can receive input operational parameters in real time and quickly output optimized ink wash effect parameters. This ensures that the digital brush accurately responds to the artist's operational intentions while presenting an ink wash effect that conforms to real physical laws, further enhancing the realism and interactivity of digital drawing.

[0053] In this embodiment, preferably, the ink wash texture includes ink stain texture and smudged edge texture. By integrating the physical properties of the ink wash texture, the penetration, accumulation, and drying process of ink on Xuan paper can be accurately simulated. By integrating the physical properties of the paper texture, the digital brush can simulate the penetration resistance and diffusion path of ink on different types of paper. The dual fusion of paper texture and ink wash texture not only accurately reproduces the texture of ink on real paper, but also allows artists to experience different styles of smudged effects in real time, from the delicate brushwork of sized Xuan paper to the freehand brushwork of unsized Xuan paper, by switching paper texture parameters during digital painting. This further enhances the style adaptability and immersive experience of the digital brush. The color blending algorithm employs either a physically-based layered overlay algorithm or a nonlinear color interpolation algorithm. The layered overlay algorithm simulates the gradual penetration of ink into the fibers of Xuan paper, dividing each stroke into a base layer, a transition layer, and a surface layer. It adjusts the density and transparency of each layer to achieve a natural sense of depth from dark to light. The nonlinear color interpolation algorithm addresses the color gradation requirements at the edges of the blending effect. By dynamically adjusting the curvature of the interpolation function, it smoothly transitions the color values ​​of adjacent pixels, avoiding the harsh boundaries caused by traditional linear interpolation and improving blending efficiency. The strong blending of softness and realism at the edges combines the physical properties of ink and paper textures, seamlessly connecting the gradual penetration of ink, edge gradation, and natural color transition. This allows each stroke of the digital brush to possess both the texture of traditional ink and the fluidity of blending. Furthermore, through parametric adjustments, artists can control the diffusion speed, ink tones, and edge sharpness in real time, achieving a stylistic transformation from the delicate blending of sized Xuan paper to the unrestrained diffusion of unsized Xuan paper. This truly achieves accurate reproduction and creative expression of traditional Chinese ink painting effects in digital media.

[0054] In this embodiment, preferably, the general interface uses a plug-in interface that supports mainstream digital painting software such as Photoshop and Procreate. This allows for seamless integration of the calculation results of layer overlay algorithms and nonlinear color interpolation algorithms into the software's native brush rendering pipeline. Through the software's familiar brush settings interface, parameters such as the diffusion speed of ink wash, the distribution of ink tones, and the degree of edge sharpness can be adjusted in real time without switching to a separate tool or learning new operating logic. This enables the direct style conversion between delicate ink wash on sized Xuan paper and free-flowing ink wash on raw Xuan paper within Photoshop's layer workflow or Procreate's gesture operations, making the creation process of digital ink painting more in line with the usage habits of traditional artists.

[0055] In this embodiment, the preferred subjective evaluation includes inviting professional traditional Chinese painters to rate the technique reproduction and artistic expression of the ink wash effect on a ten-point scale, and collecting feedback from ordinary users on the operation response speed, effect controllability, and interface friendliness through online questionnaires. Objective indicators include the smoothness of the ink wash edge (by calculating the standard deviation of the edge pixel color gradient, the smaller the value, the softer the edge), the richness of the ink color layering (by counting the number of effective peaks in the color histogram, the more peaks and the more evenly distributed, the stronger the layering), and the frame rate stability of real-time rendering (at 1080P resolution and the highest pressure sensitivity level, the frame rate fluctuation should not exceed 5fps within 5 minutes of continuous drawing). That is, by combining subjective evaluation with objective indicators, the comprehensive performance of the dynamic digital brush in terms of the reproduction of traditional Chinese ink wash effect, creative controllability, and real-time performance can be verified. This ensures that it not only meets the stringent requirements of professional traditional Chinese painters for accurate reproduction of traditional brush and ink techniques, but also meets the expectations of ordinary digital painting enthusiasts for ease of operation and controllable effects, thus providing a more scientific evaluation basis for the promotion of digital ink painting and the digital inheritance of traditional techniques.

[0056] This invention constructs a multi-dimensional blurring effect generation model, incorporating structural parameters such as the fiber density and porosity of Xuan paper, physical properties such as ink concentration and permeability coefficient, and traditional techniques such as multi-layer stacking in ink accumulation and dry-wet interleaving in ink breaking. This model accurately reproduces the dynamic processes of ink penetration, diffusion, evaporation, and multi-layer stacking on Xuan paper, achieving the textures of traditional brush and ink, such as the gradual transition from light to dark and the collision effect of dry brush and wet paper. It avoids common problems in digital rendering, such as flat washes and harsh boundaries, giving the ink both a heavy, layered feel and a dynamic, blurring effect. Furthermore, through dynamic parameter mapping and a real-time rendering engine, the user's pen pressure intensity, stroke speed, and brush tip angle are transformed in real-time into blurring effect parameters such as diffusion range, concentration gradient, and edge blurring degree. Combined with GPU-accelerated ink texture synthesis, paper texture mapping and layering, and non-linear color interpolation algorithms, the blurring effect is realized in real time. Real-time feedback—users can instantly see the ink spreading along the brushstrokes and the water penetrating into the gaps between the Xuan paper fibers, just like the natural feedback of a real brush painting on Xuan paper. The operation feels closer to that of a traditional brush, allowing even veteran artists accustomed to traditional painting to quickly adapt and precisely control the diffusion effect. At the same time, through iterative parameter optimization based on user feedback (such as adjusting the mapping curve between pen pressure and ink density, and the response threshold between speed and diffusion range) and universal interfaces supporting mainstream software such as Photoshop and Procreate, the sensitivity of the operation response and software compatibility have been further improved. It supports style transitions from delicate diffusion on sized Xuan paper to unrestrained diffusion on unsized Xuan paper, satisfying the needs of professional creators for precise control of diffusion effects (such as high-resolution painting of large landscape scrolls and delicate brushstrokes in flower and bird painting), while also providing ordinary enthusiasts with convenient digital creation tools (brushes can be accessed without learning new operating logic). In summary, this invention not only achieves a resemblance in form to the traditional ink wash effect (such as ink tones, blurred edges, and fibrous texture), but also a resemblance in spirit (such as the freehand quality of broken brushstrokes and the realism of the tactile experience). It provides strong technical support for the digital inheritance and creative expression of traditional Chinese painting, making digital media an extension rather than a replacement for traditional Chinese painting, and promoting the deep integration of traditional art and modern technology.

[0057] Please see Figure 2 The present invention also provides a dynamic digital brush generation system for achieving the ink wash effect in traditional Chinese painting, comprising:

[0058] Ink wash model building unit: used to collect the physical parameters of Xuan paper and the property parameters of ink, integrate traditional techniques and rules, and build a physical-technical fusion model for generating wash effects;

[0059] Dynamic parameter mapping unit: used to establish the mapping relationship between user operation parameters and blurring effect parameters, and realize real-time dynamic response through machine learning model;

[0060] Real-time rendering unit: Used for real-time rendering of ink wash texture synthesis, paper texture mapping, and color mixing based on GPU acceleration, generating natural blurring effects;

[0061] Interactive interface unit: Based on a touch screen, it provides an intuitive parameter adjustment interface, allowing users to quickly set the blurring effect parameters and preview the effect in real time;

[0062] Compatibility adapter unit: Used to achieve system compatibility with different digital painting software through a universal interface, thereby expanding the application scope of the system.

[0063] In this embodiment, preferably, the ink wash model construction unit includes:

[0064] Xuan paper physical parameter acquisition submodule: By analyzing the microstructural characteristics of Xuan paper fiber density and interlacing morphology through high-resolution microscopic images, and combined with ink absorption experiments under constant temperature and humidity environment, parameters such as ink absorption rate, diffusion coefficient and water evaporation rate are measured to establish a digital physical model of Xuan paper material, so as to achieve accurate simulation of ink absorption characteristics of different types of Xuan paper.

[0065] The ink property parameter acquisition submodule uses a spectrometer to obtain the spectral reflectance of ink at different dilution ratios, converts it into quantized values ​​in RGB and HSV color spaces, measures the dynamic viscosity of ink using a rotational viscometer, and obtains the diffusion coefficient of ink in liquid by combining diffusion experiments. It constructs a correlation model between ink concentration, viscosity and diffusion characteristics, and supports accurate reproduction of five ink color levels: "burnt, thick, heavy, light and clear".

[0066] The traditional techniques and rules integration submodule extracts the rule parameters of core techniques such as ink accumulation (the number of times and intervals of multiple layers of ink), ink breaking (the ratio of ink volume and the order of layering when dry and wet inks interact), and ink blending (the trajectory and pause time of brush strokes) by motion capture and interviews with traditional Chinese painters on their ink blending operations. These parameters are then transformed into algorithmic logic based on probability statistics or machine learning and are used in real-time collaborative calculations with the physical model of Xuan paper and the ink and water property model to form a dual-driven ink blending effect generation mechanism that combines physical characteristics and technique rules.

[0067] Through the synergistic cooperation of the above sub-modules, the ink wash model building unit can accurately integrate the material characteristics of Xuan paper, the physical and chemical properties of ink, and the rules of traditional ink wash techniques, forming an ink wash model that combines physical realism and artistic expression. This provides core underlying logic support for dynamic digital brushes to achieve ink wash effects that conform to the rules of traditional painting.

[0068] In this embodiment, preferably, the dynamic parameter mapping unit includes:

[0069] The user operation parameter perception submodule uses a digital tablet pressure sensor, an intelligent pen posture capture device, and a user interface to collect dynamic operation parameters in real time, such as the peak pressure, pen speed change rate, pen tip tilt angle, pen stroke continuity, and dwell time. At the same time, it acquires static parameters preset by the user, such as the type of Xuan paper, ink concentration, and blending method (e.g., dry blending, wet blending). It normalizes multi-source heterogeneous data into a unified format operation parameter matrix to achieve precise quantification of user intent.

[0070] The sub-module for correlation of ink diffusion effect parameters: Based on the effect parameters such as diffusion radius, ink gradient distribution coefficient, blurring boundary ambiguity, and ink layer thickness generated by the ink diffusion model, the module mines the potential correlation between user operation parameters and ink diffusion effect parameters through mutual information entropy calculation or nonlinear principal component analysis, establishes an operation and effect correlation feature library, and uses a support vector machine classifier to label feature weights to distinguish the degree of influence of core operation parameters and secondary operation parameters on the effect.

[0071] The real-time mapping model construction submodule uses a lightweight Transformer model combined with a deep reinforcement learning framework to build a real-time mapping model. It takes the user operation parameter matrix as input and the shading effect association features as output. The input sample data is pre-trained, and the model parameters are updated in real time through an online learning mechanism to achieve the effect parameter response within 50 milliseconds when the user operation changes. This ensures that the shading effect is presented naturally with the dynamic adjustment of the user's pen movements, achieving an interactive experience where the pen moves with the heart.

[0072] Through the coordinated operation of the aforementioned sub-modules, the dynamic parameter mapping unit can accurately realize the dynamic conversion and real-time output of parameters from user brush strokes to ink wash effect parameters, providing high-fidelity effect parameter input for the subsequent real-time rendering module of digital brushes. The closed-loop collaborative mechanism from user operation perception to effect feature association to real-time model mapping ensures that the ink wash effect transitions naturally with subtle changes in the user's brush pressure, brush speed fluctuations, or brush tip tilt angle adjustments. It not only retains the sense of layering and randomness of the ink wash effect in traditional Chinese painting, but also achieves a high degree of controllability of operation and effect through digital technology. It truly becomes the core bridge connecting the user's painting intentions and the dynamic rendering of digital brushes, laying a solid parameter mapping foundation for the entire system to achieve dynamic ink wash effects that follow the user's wishes and create natural ink wash effects.

[0073] In this embodiment, preferably, the real-time rendering unit includes:

[0074] The ink wash texture synthesis submodule utilizes a GPU-based CUDA / Metal parallel computing framework. It combines parameters such as the diffusion radius and ink gradient distribution coefficient output from the ink wash model with conditions for generative adversarial networks to pre-train an ink wash texture generation model. This generates ink wash textures that match the current diffusion state in real time. The model uses diffusion effect parameters as input and generates pixel-level ink density distributions through GPU-accelerated depth convolution operations. Worley noise is introduced to simulate the random resistance effect of Xuan paper fibers on ink, enhancing the natural randomness of the texture and ensuring a rendering frame rate of over 30fps at 4K resolution, meeting the real-time requirements of professional creation.

[0075] Paper Texture Mapping Submodule: Utilizing the digital model of fiber density and interlacing morphology constructed by the Xuan paper physical parameter acquisition submodule, the geometric texture of Xuan paper fibers is generated through 3D point cloud reconstruction technology and converted into normal maps (reflecting fiber concavity and convexity) and displacement maps (simulating fiber thickness). Through the GPU texture mapping pipeline, the paper texture and ink texture are fused in multiple channels to simulate the difference in fiber adsorption of ink color. At the same time, the diffuse reflection and specular reflection characteristics of the paper surface are adjusted through a micro-surface-based BRDF (bidirectional reflectance distribution function) model to restore the texture of Xuan paper that is "moist but not dry after being dyed with pine soot ink".

[0076] The color mixing calculation submodule is based on traditional techniques (such as the number of times ink is layered and the order of dry and wet ink) and water-ink property models (concentration, viscosity). It uses the Beer-Lambert light absorption law combined with the Kubelka-Munk theory to construct a physical color mixing model, which calculates the superposition effect of multiple layers of ink in real time. The model converts each layer of ink into the Lab color space and performs non-linear mixing according to the technique of dry ink base and wet ink glazing. At the same time, it dynamically adjusts the mixing weights based on the ink absorption rate of Xuan paper and the water evaporation parameters (the mixing weight of areas with fast ink absorption is increased by 1.2 to 1.8 times). It also uses the fragment shader of the GPU to realize pixel-by-pixel color calculation, ensuring that the transition of ink from dark to light is natural and smooth, avoiding color block breaks or color overflow, and truly restoring the layered effect of ink with five colors in traditional ink blending.

[0077] Through the coordinated operation of the aforementioned sub-modules, the real-time rendering unit is able to generate ink wash effects in digital painting scenes that highly conform to the rules of traditional Chinese painting.

[0078] The ink-wash diffusion model construction unit collects physical parameters such as the density of Xuan paper fibers and the ink absorption rate, combines them with properties such as ink concentration and viscosity, and integrates traditional technique rules such as the number of ink layers and the order of ink breaking to construct a diffusion model driven by both physical characteristics and technique logic. This achieves precise coordination between the adsorption and inhibition of ink by Xuan paper fibers, the diffusion and penetration of ink within the paper, and traditional technique movements. It solves the problem of existing technologies that rely solely on simple texture superposition or single physical simulation, resulting in a lack of depth and traditional charm in the diffusion effect. The dynamic parameter mapping unit uses a digital tablet pressure sensor and posture capture device to perceive user pen pressure peaks, pen speed change rates, and other operation parameters in real time. It uses mutual information entropy calculation to explore the potential correlation between operation and effect parameters such as diffusion radius and ink color gradient, and adopts a lightweight Transformer model to achieve 50 milliseconds. The system's real-time mapping addresses the pain point of "the idea arrives but the brush doesn't" caused by high latency and weak correlation between operation and effect. The real-time rendering unit synthesizes ink diffusion textures using GPU-accelerated conditional generative adversarial networks, combining Xuan paper fiber normal maps and displacement maps to simulate paper texture. It employs Beer-Lambert's law and Kubelka-Munk theory to achieve non-linear ink mixing, ensuring a rendering frame rate of over 30fps at 4K resolution and a natural transition of ink from dark to clear, solving the problems of rendering latency and ink color banding in high-resolution scenes. The compatibility adaptation unit, through a universal interface design, supports seamless integration with mainstream digital painting software such as Photoshop, SAI, and Procreate, overcoming the limitation of the system's inability to be widely applied in professional creative scenarios due to compatibility issues. Through a full-chain design encompassing physical model construction, dynamic parameter mapping, real-time rendering optimization, and multi-software compatibility, this system truly achieves the digital dynamic restoration of traditional Chinese ink painting diffusion effects, providing digital artists with a tool that combines the expressive power of traditional art with the convenience of modern creation, effectively promoting the integration and innovation of traditional Chinese painting techniques and digital media technology.

[0079] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for generating dynamic digital brushes to achieve the ink wash effect in traditional Chinese painting, characterized in that, Includes the following steps: S1. Construct a multi-dimensional model for generating ink diffusion effects: Collect the fiber structure parameters of traditional Xuan paper and the physical properties of ink, establish a simulation model of ink diffusion in Xuan paper, extract the rules of traditional ink diffusion techniques, integrate physical simulation with the rules of techniques, and establish a multi-dimensional model for generating ink diffusion effects. S2. Dynamic Parameter and Blur Effect Mapping: Establish a mapping relationship between user operation parameters and blur effect parameters, train the mapping model through machine learning algorithms, and realize the real-time dynamic response of user operation to blur effect; S3. Design a real-time rendering engine: Based on GPU acceleration technology, it performs real-time rendering and integrates ink texture synthesis, paper texture mapping, and color mixing algorithms. During the real-time rendering process, it combines the blurring effect generation model and dynamic parameter mapping results. According to different user operations and preset blurring effect requirements, it calculates the color value and texture information of each pixel to realize the real-time generation and display of the blurring effect. S4. Parameter Optimization and Compatibility Adaptation: Iteratively optimize parameter mapping relationships through user feedback to improve the sensitivity and finesse of operation response; establish a universal interface to achieve compatibility adaptation between the system and different software. S5. Effect Evaluation and Continuous Iteration: Construct a multi-dimensional blending effect evaluation system, combine subjective evaluation and objective indicators to verify the authenticity and artistry of the generated effect, and continuously iterate and optimize the model and parameters based on the evaluation results to improve the system's generation quality and user experience.

2. The method for generating dynamic digital brushes to achieve the ink wash effect in traditional Chinese painting, as described in claim 1, is characterized in that... The fiber structure parameters include fiber density and porosity; the physical properties of the ink include ink concentration and permeability coefficient; the rules of ink blending techniques include multi-layer superposition of ink accumulation method and dry and wet breaking of ink method.

3. The method for generating dynamic digital brushes to achieve the ink wash effect in traditional Chinese painting, as described in claim 1, is characterized in that... The operating parameters include pressure sensitivity, pen stroke speed, and pen tip angle; the blurring effect parameters include diffusion range, density gradient, and edge blurring degree; the machine learning algorithm is either a support vector machine or a neural network.

4. The method for generating dynamic digital brushes to achieve the ink wash effect in traditional Chinese painting, as described in claim 1, is characterized in that... The ink texture includes ink stain texture and smudged edge texture; the paper texture includes the smooth texture of sized Xuan paper and the rough texture of unsized Xuan paper; the color mixing algorithm uses a physics-based layered overlay algorithm or a nonlinear color interpolation algorithm. The layered overlay algorithm simulates the process of ink penetrating layer by layer between Xuan paper fibers, dividing the ink of each stroke into a base layer, a transition layer, and a surface layer, and adjusting the concentration and transparency of each layer to achieve a natural sense of layering from dark to light. The nonlinear color interpolation algorithm addresses the color gradation requirements of smudged edges by dynamically adjusting the curvature of the interpolation function to smoothly transition the color values ​​of adjacent pixels, avoiding the harsh boundaries caused by traditional linear interpolation and enhancing the softness and realism of the smudged edges.

5. The method for generating dynamic digital brushes to achieve the ink wash effect in traditional Chinese painting, as described in claim 1, is characterized in that... The general interface uses a plug-in interface that supports mainstream digital painting software such as Photoshop and Procreate.

6. The method for generating dynamic digital brushes to achieve the ink wash effect in traditional Chinese painting, as described in claim 1, is characterized in that... The subjective evaluation includes inviting professional Chinese painters to score the technique reproduction and artistic expression of the blending effect on a ten-point scale, and collecting feedback from ordinary users on the operation response speed, effect controllability, and interface friendliness through online questionnaires; the objective indicators include the smoothness of the blending edge, the richness of the ink color layering, and the frame rate stability of real-time rendering.

7. A dynamic digital brush generation system for achieving the ink wash effect in traditional Chinese painting, characterized in that, include: Ink wash model building unit: used to collect the physical parameters of Xuan paper and the property parameters of ink, integrate traditional techniques and rules, and build a physical-technical fusion model for generating wash effects; Dynamic parameter mapping unit: used to establish the mapping relationship between user operation parameters and blurring effect parameters, and realize real-time dynamic response through machine learning model; Real-time rendering unit: Used for real-time rendering of ink wash texture synthesis, paper texture mapping, and color mixing based on GPU acceleration, generating natural blurring effects; Interactive interface unit: Based on a touch screen, it provides an intuitive parameter adjustment interface, allowing users to quickly set the blurring effect parameters and preview the effect in real time; Compatibility adapter unit: Used to achieve system compatibility with different digital painting software through a universal interface, thereby expanding the application scope of the system.

8. The dynamic digital brush generation system for achieving the ink wash effect in traditional Chinese painting, as described in claim 7, is characterized in that... The ink wash model construction unit includes: Xuan paper physical parameter acquisition submodule: By analyzing the microstructural characteristics of Xuan paper fiber density and interlacing morphology through high-resolution microscopic images, and combined with ink absorption experiments under constant temperature and humidity environment, parameters such as ink absorption rate, diffusion coefficient and water evaporation rate are measured to establish a digital physical model of Xuan paper material, so as to achieve accurate simulation of ink absorption characteristics of different types of Xuan paper. The ink property parameter acquisition submodule uses a spectrometer to obtain the spectral reflectance of ink at different dilution ratios, converts it into quantized values ​​in RGB and HSV color spaces, measures the dynamic viscosity of ink using a rotational viscometer, and obtains the diffusion coefficient of ink in liquid by combining diffusion experiments. It constructs a correlation model between ink concentration, viscosity and diffusion characteristics, and supports accurate reproduction of the five ink color levels of "burnt, thick, heavy, light and clear". The traditional techniques and rules integration submodule extracts the rule parameters of the core techniques of ink accumulation, ink breaking, and ink blending through motion capture and interviews with traditional Chinese painters. These parameters are then transformed into algorithmic logic based on probability statistics or machine learning. This logic is then used in real-time collaborative calculations with the physical model of Xuan paper and the ink property model to form a dual-driven ink blending effect generation mechanism that combines physical properties and technique rules.

9. A dynamic digital brush generation system for achieving ink wash effect in traditional Chinese painting, as described in claim 7, is characterized in that... The dynamic parameter mapping unit includes: The user operation parameter perception submodule uses a digital tablet pressure sensor, an intelligent pen posture capture device, and a user interface to collect dynamic operation parameters in real time, such as the peak pressure, pen speed change rate, pen tip tilt angle, pen stroke continuity, and dwell time. At the same time, it acquires static parameters preset by the user, such as the type of Xuan paper, ink concentration, and blending method. It normalizes multi-source heterogeneous data into a unified format operation parameter matrix to achieve accurate quantification of user intent. The sub-module for correlation of ink diffusion effect parameters: Based on the effect parameters such as diffusion radius, ink gradient distribution coefficient, blurring boundary ambiguity, and ink layer thickness generated by the ink diffusion model, the module mines the potential correlation between user operation parameters and ink diffusion effect parameters through mutual information entropy calculation or nonlinear principal component analysis, establishes an operation and effect correlation feature library, and uses a support vector machine classifier to label feature weights to distinguish the degree of influence of core operation parameters and secondary operation parameters on the effect. The real-time mapping model construction submodule uses a lightweight Transformer model combined with a deep reinforcement learning framework to build a real-time mapping model. It takes the user operation parameter matrix as input and the shading effect association features as output. The input sample data is pre-trained, and the model parameters are updated in real time through an online learning mechanism to achieve the effect parameter response within 50 milliseconds when the user operation changes. This ensures that the shading effect is presented naturally with the dynamic adjustment of the user's pen movements, achieving an interactive experience where the pen moves with the heart.

10. A dynamic digital brush generation system for achieving the ink wash effect in traditional Chinese painting, as described in claim 7, is characterized in that... The real-time rendering unit includes: The ink wash texture synthesis submodule utilizes a GPU-based CUDA / Metal parallel computing framework. It combines parameters such as the diffusion radius and ink gradient distribution coefficient output from the ink wash model with conditions for generative adversarial networks to pre-train an ink wash texture generation model. This generates ink wash textures that match the current diffusion state in real time. The model uses diffusion effect parameters as input and generates pixel-level ink density distributions through GPU-accelerated depth convolution operations. Worley noise is introduced to simulate the random resistance effect of Xuan paper fibers on ink, enhancing the natural randomness of the texture and ensuring a rendering frame rate of over 30fps at 4K resolution, meeting the real-time requirements of professional creation. Paper Texture Mapping Submodule: Utilizing the digital model of fiber density and interlacing morphology constructed by the Xuan paper physical parameter acquisition submodule, the geometric texture of Xuan paper fibers is generated through 3D point cloud reconstruction technology and converted into normal maps and displacement maps; through the GPU texture mapping pipeline, the paper texture and ink texture are fused in multiple channels to simulate the difference in fiber adsorption of ink color, and at the same time, the diffuse reflection and specular reflection characteristics of the paper surface are adjusted through a micro-surface-based BRDF model to restore the texture of Xuan paper; The color mixing calculation submodule, based on traditional techniques and ink properties, uses the Beer-Lambert light absorption law combined with Kubelka-Munk theory to construct a physical color mixing model, which calculates the superposition effect of multiple ink layers in real time. The model converts each layer of ink into the Lab color space and performs non-linear mixing according to the technique of dry ink base and wet ink glazing. At the same time, it dynamically adjusts the mixing weight by combining the ink absorption rate of Xuan paper and the water evaporation parameters, and realizes pixel-by-pixel color calculation through the fragment shader of the GPU to ensure that the transition of ink from dark to light is natural and smooth, avoiding color block breaks or color overflow, and truly restoring the layered effect of ink with five colors in traditional ink wash.

Images