Facial image generation system, apparatus and device based on parameterized style framework

Abstract

The application relates to a face image generation system, device and equipment based on a parameterized style framework, and a method comprising the following steps: based on a face shape and an aesthetic framework, a quantitative constraint parameter is obtained by decomposing a face feature acquisition amount through a vertex coordinate offset amount; a constraint parameter set is generated; according to the constraint parameter set, a fixed parameter combination is allocated for a specific style, and a preset face decoration is bound in association to form a preset face kneading template; according to a plurality of part presets selected by a user, a parameter is automatically decomposed and fused through a weight conflict resolution rule to generate a mixed style parameter value range, so that style diversification and controllability in a unified aesthetic framework are realized. Through the construction of the constraint parameter set, the binding of the style preset and the resolution of the parameter conflict, one-key generation and efficient face kneading with diversified personalization are realized under the premise of ensuring the uniformity of the style.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a facial image generation system, apparatus, and device based on a parametric style framework. Background Technology

[0002] In the field of game character customization, the facial creation system, as a core function to enhance player immersion and character identification, directly impacts user experience and the consistency of game art style. With players' increasing demand for personalized appearances and rising aesthetic standards, game products urgently need efficient and stable solutions to generate character facial images that possess both unique personality and distinctive style. However, traditional facial creation methods exhibit significant limitations when faced with stylistic requirements with clear aesthetic guidelines, such as "ancient style" or "cyberpunk," failing to guarantee the artistic quality and stylistic consistency of the generated images while granting users freedom.

[0003] Currently, most character creation systems remain at the stage of providing broad, undefined adjustments to basic parameters, such as boundless sliders for controlling facial features and contour curvature. Such systems fail to effectively convey specific stylistic aesthetic concepts, forcing players to rely on extensive trial-and-error adjustments based on personal experience and intuition. This not only results in low operational efficiency but also often leads to ambiguous stylistic orientations, mixed features, and insufficient recognizability. Furthermore, the lack of unified stylistic guidance and quantitative standards makes the customization process among different players with similar aesthetic intentions highly prone to convergence. A large number of character designs are concentrated in a few "internet celebrity face" paradigms, exacerbating homogenization and contradicting the original intention of character creation systems to encourage individuality.

[0004] Therefore, in the field of character creation technology, how to build a system that can transform abstract aesthetic frameworks into quantifiable and executable parameter rules, thereby achieving efficient generation of personalized characters while ensuring stylistic consistency and artistry, has become a key technical problem that urgently needs to be solved. Summary of the Invention

[0005] This application provides a facial image generation system based on a parametric style framework, characterized by comprising: Based on the facial shape and aesthetic framework, facial features are decomposed by vertex coordinate offset to obtain quantified constraint parameters and generate a constraint parameter set; Based on the set of constraint parameters, a preset face-shaping template is formed by assigning a fixed combination of parameters to a specific style and binding it with a preset association with facial decorations. Based on multiple preset parts selected by the user, the parameters are automatically decomposed and fused according to the weight conflict resolution rules to generate a range of mixed style parameter values, thereby achieving style diversity and controllability within a unified aesthetic framework.

[0006] Optionally, the step of using facial shape and aesthetic framework to decompose facial features through vertex coordinate offsets to obtain quantized constraint parameters and generate a constraint parameter set includes: Based on the preset aesthetic framework requirements, a series of quantifiable adjustment dimensions are obtained by decomposing facial features into components to be adjusted corresponding to the key vertex set on the basic 3D model. By using vertex coordinate offset as the sole quantification standard and setting offset thresholds that match aesthetic boundaries for the displacement of each vertex set on the X, Y, and Z axes, a set of constraint parameters with controlled ranges of all parameter values ​​is generated.

[0007] Optionally, the step of forming a preset face-shaping template by assigning fixed parameter combinations to a specific style and binding them with preset facial decorations according to the constraint parameter set includes: Based on the generated set of constraint parameters, the core facial feature presets for a specific style are formed by assigning fixed parameter combinations to a specific style. By logically associating and packaging the core facial feature presets with the facial decoration presets specified in the system resource library at the data layer, a preset face-shaping template is obtained.

[0008] Optionally, the step of automatically decomposing and merging parameters based on multiple preset parts selected by the user, using weight conflict resolution rules, to generate a range of mixed style parameter values, thereby achieving style diversity and controllability within a unified aesthetic framework, includes: Based on the user's selection of multiple preset operation commands for various parts, the system automatically decomposes the specific feature parameters corresponding to these presets and obtains multiple sets of parameter data from different sources; By using preset weight conflict resolution rules, the mixed parameter data is fused and calculated to generate the value range of the mixed style parameters.

[0009] Optionally, by using vertex coordinate offset as the sole quantization standard and setting offset thresholds matching aesthetic boundaries for the displacement of each vertex set on the X, Y, and Z axes, a set of constraint parameters whose value ranges are all controlled is generated, including: By mapping the style feature dimension to the positive displacement of a specific set of key vertices in the axial coordinates of the basic 3D model, the quantification of facial features is achieved. By setting a closed interval for the displacement of each set of vertex coordinates, which is strictly limited by the minimum and maximum offset thresholds, a set of constraint parameters that prevents style deviation and is mathematically strictly controlled is generated.

[0010] Optionally, the step of obtaining a preset face-shaping template by logically associating and packaging the core facial feature presets with the facial decoration presets specified in the system resource library at the data layer includes: By establishing an inseparable data relationship between the fixed parameter combination representing the shape of the facial features in the database and the facial decoration preset representing the appearance, the inherent binding logic between style elements is constructed. By encapsulating all these associated style elements into data entities that can be directly called, you can obtain preset face-shaping templates that are ready to use and have a highly coordinated and unified presentation effect.

[0011] Optionally, the step of performing fusion calculations on the mixed parameter data using preset weight conflict resolution rules to generate the value range of the mixed style parameters includes: By establishing predefined weight conflict resolution rules, the logical basis for the system to make automatic decisions is established; By applying the aforementioned weight conflict resolution rules to traverse and merge all conflict parameters, a range of mixed-style parameter values ​​that is subject to strict mathematical constraints is generated.

[0012] This application also provides a facial image generation device based on a parametric style framework, characterized in that the device comprises: The parameterized framework construction module is used to decompose facial features based on facial shape and aesthetic framework, obtain quantized constraint parameters by vertex coordinate offset, and generate a set of constraint parameters. The preset encapsulation and binding module is used to form a preset face-shaping template by assigning a fixed combination of parameters to a specific style and binding it with the preset facial decoration according to the constraint parameter set. The style fusion calculation module is used to automatically decompose parameters and perform fusion calculations based on multiple preset parts selected by the user, through weight conflict resolution rules, to generate a range of mixed style parameter values, thereby achieving style diversity and controllability within a unified aesthetic framework.

[0013] Optionally, the style fusion calculation module further includes: The parameter decomposition and acquisition module is used to automatically parse and extract the underlying parameter data based on multiple preset parts selected by the user. The weighting and conflict arbitration module is used to perform weighted calculations and arbitrations on conflicting parameters from different presets according to preset weight conflict resolution rules. The fusion range generation module is used to calculate and output the new legal value range of each parameter under the fusion style based on the result of conflict arbitration.

[0014] This application also provides an electronic device, characterized in that it is used to implement any of the aforementioned face image generation systems based on a parametric style framework, including... The processor is used to perform all computational tasks to implement a face generation system based on a parametric style framework; Memory is used to store processor-executable instructions and statically stored data.

[0015] The beneficial effects of this application are as follows: By decomposing facial features into a set of constraint parameters based on original facial images and a target aesthetic framework, with vertex coordinate offset as the quantification standard, and setting a value range for each parameter that strictly matches the aesthetic boundary, a strong constraint on the generated image style is achieved from the mathematical level, fundamentally solving the problem of style deviation and image collapse caused by the lack of boundary adjustment in traditional free face sculpting. By assigning fixed parameter combinations to specific styles according to the constraint parameter set and binding them with preset facial decorations such as hairstyles and makeup at the data layer, a ready-to-use preset face sculpting template is formed, realizing one-click generation of a highly coordinated and stylistically unified complete image, greatly reducing the user's operation threshold and time cost. Furthermore, when the user selects multiple preset parts, the system automatically decomposes the parameters and performs fusion calculations according to predefined weight conflict resolution rules, generating a new and controlled parameter value range under the mixed style. This enables the intelligent and controllable generation of a rich variety of fusion styles within a unified aesthetic framework, breaking through the style limitations of a single preset and achieving a balance between personalization and aesthetic taste. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings required in the description of the embodiments or the prior art are briefly introduced below. Obviously, the accompanying drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 A flowchart illustrating a specific embodiment of the face image generation system based on a parametric style framework of this application is shown. Figure 2 This diagram illustrates a face image generation system based on a parametric style framework according to a specific embodiment of this application. Figure 3 This is a block diagram of a face image generation apparatus based on a parametric style framework according to a specific embodiment of this application. Detailed Implementation

[0018] Various exemplary embodiments, features, and aspects of this application will now be described in detail with reference to the accompanying drawings. The same reference numerals in the drawings denote elements that have the same or similar functions. Although various aspects of the embodiments are shown in the drawings, they are not necessarily drawn to scale unless specifically indicated otherwise.

[0019] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0020] The term “exemplary” as used herein means “serving as an example, embodiment, or illustration.” Any embodiment illustrated herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.

[0021] Furthermore, to better illustrate this application, numerous specific details are provided in the following detailed embodiments. Those skilled in the art should understand that this application can be implemented without certain specific details. In some instances, methods, means, components, and circuits well-known to those skilled in the art have not been described in detail in order to highlight the main points of this application.

[0022] This application proposes an intelligent face-shaping method and system based on a parametric style framework to address the technical challenge of balancing inconsistent styles with user-defined freedom in digital human creation. Based on a clear aesthetic framework, this application proposes a constraint parameter set construction scheme using vertex coordinate offset as the core quantification standard. By defining key vertex sets and their coordinate offset thresholds, abstract aesthetics are transformed into precise and executable mathematical rules from the bottom layer of the 3D model. Based on this, this application designs a preset encapsulation and binding mechanism and a style fusion calculation process. The former generates a preset template with a highly consistent style by associating fixed parameter combinations with decorative resources; the latter intelligently decomposes and fuses multiple preset parameters selected by the user through the introduction of weight conflict resolution rules, dynamically generating a new parameter range that conforms to aesthetic boundaries. While ensuring that all generated images do not deviate from the target style framework, this application not only significantly reduces the operational burden on users in creating high-quality images but also achieves stylistic diversity and originality, providing a reliable technical foundation for the large-scale, standardized, and personalized generation of digital images.

[0023] Example 1 like Figure 1 The diagram shown is a flowchart of a face image generation system based on a parametric style framework according to an embodiment of this application, specifically including the following: S100, based on a facial shape and aesthetic framework, decomposes facial features by vertex coordinate offset to obtain quantized constraint parameters and generate a constraint parameter set.

[0024] Specifically, based on a pre-defined aesthetic framework and a basic facial model, a quantifiable style constraint system is established. The target aesthetic style is deconstructed into a series of key facial feature dimensions, and each feature dimension is mapped to one or more key vertex sets in the basic 3D mesh model. The quantification standard for each constraint parameter is uniquely defined as the coordinate offset of the corresponding vertex set in the model space coordinate system. By setting maximum and minimum offset thresholds for each vertex set along the X, Y, and Z axes, the adjustable boundary of the parameter at the physical level is precisely defined. These thresholds are not technical extremes, but rather numerical ranges that have been aesthetically verified and strictly match the visual boundary of the target style. Finally, all such boundary-limited parameters are systematically integrated to generate a complete and closed set of constraint parameters, thereby constructing an insurmountable style protection framework at the data level.

[0025] S200, based on the set of constraint parameters, a preset face-shaping template is formed by assigning a fixed combination of parameters to a specific style and binding it with a preset association with facial decoration.

[0026] Specifically, based on the constraint parameter set generated in step S100, the system performs data encapsulation and resource binding for style presets. First, for each desired style, a qualified and fixed combination of parameters is assigned to each facial component from the constraint parameter set. This combination serves as the mathematically unique identifier of the style, ensuring that the facial features strictly adhere to the preset aesthetic standards. To achieve completeness and consistency in style expression, the system further logically associates and packages this core facial feature parameter combination with pre-designed facial decoration presets in the resource library that match the visual tone. These decoration presets cover non-geometric elements such as eyelashes, pupils, and makeup. Through this strong association binding, an originally abstract style concept is materialized into a preset face-shaping template containing all necessary rendering information, ready to use, providing users with the ability to apply a complete style experience with a single click.

[0027] S300 automatically decomposes and merges parameters based on multiple preset parts selected by the user, using weight conflict resolution rules, to generate a range of mixed style parameter values, achieving style diversity and controllability within a unified aesthetic framework.

[0028] Specifically, in response to a user's command to mix and match multiple preset body parts from the preset template library, the system performs automated style fusion calculations. This process begins with reverse analysis of each preset selected by the user. The system automatically disassembles these presets, extracting the specific quantitative parameter data corresponding to them, thereby obtaining multiple sets of raw data from different sources that may have numerical conflicts. The core of the fusion process relies on predefined weight conflict resolution rules, which constitute the logical basis for the system's intelligent decision-making and are used to arbitrate competing parameter values ​​between different presets. Based on these rules, the system iterates through and fuses all conflicting parameters. The output is not a single fixed value, but a dynamically generated new range of parameter values. This newly generated range is also constrained by the global aesthetic framework, thus ensuring at the algorithmic level that all possible facial images generated can satisfy the user's personalized mixing intentions while maintaining a unified style tone, achieving a dialectical unity between diversity creation and style controllability.

[0029] In summary, this application overcomes the limitations of traditional face-shaping systems, which rely on subjective user aesthetics leading to inconsistent styles and high technical barriers. It constructs an intelligent image generation system based on a parametric aesthetic framework. First, addressing the core pain point of boundless style creation and the resulting incongruous images, a strongly constrained parameter set with vertex coordinate offsets as the quantification benchmark is established. By deconstructing abstract aesthetics into specific facial feature dimensions and mapping them to a set of key vertices in a 3D mesh, a strict three-axis displacement threshold matching the style boundary is set for each parameter. This achieves style solidification and protection of the generated image from the mathematical model's underlying layer, ensuring absolute uniformity and stability of the aesthetic output. Second, to lower the user's operational threshold and provide a complete style experience, a data encapsulation and resource binding mechanism for preset templates is designed. By assigning fixed parameter combinations to specific styles from the constraint parameter set, the mathematical uniqueness of their facial features is ensured. This core parameter is logically linked and packaged with decorative resources such as eyelashes and makeup that match the visual tone, materializing the abstract style concept into a ready-to-use complete image template, achieving one-click generation of high-quality style images. Finally, to achieve an effective balance between personalized customization and stylistic consistency, a style fusion calculation process based on intelligent decision-making was constructed. The system can automatically parse multiple preset parameters selected by the user and, according to predefined weight conflict resolution rules, arbitrate and fuse conflicting parameters, dynamically generating a new parameter value range that conforms to the global aesthetic framework. This process, at the algorithmic level, ensures that users can freely combine and explore within the established style boundaries, ultimately achieving a unity of diversity and controllability in image generation.

[0030] As an optional implementation of this application, optionally, in step S100, based on the facial shape and aesthetic framework, facial features are decomposed using vertex coordinate offsets to obtain quantized constraint parameters, generating a constraint parameter set, including: S101, based on the preset aesthetic framework requirements, obtains a series of quantifiable adjustment dimensions by decomposing facial features into components to be adjusted corresponding to the key vertex set on the basic 3D model.

[0031] Specifically, during system initialization, based on the pre-defined aesthetic framework's normative requirements for facial morphology, the system initiates a topological structure analysis process on the basic 3D facial model, transforming abstract aesthetic principles into a set of geometric elements that can be recognized and processed by the computer. The system first deconstructs the complete facial shape into multiple interconnected yet independently controllable feature components based on facial anatomical features and aesthetic compositional rules. Taking the "nose" shape as an example, the system can further refine it into three core feature dimensions: "nose bridge shape," "nose tip contour," and "nose wing width." Each defined component to be adjusted is precisely mapped to a specific set of key vertices on the surface of the basic 3D model. To achieve quantitative control over the nose dimensions, the system selects three sets of key vertices: for "nose bridge shape," the root point, midpoint, and tip of the nose are selected; for "nose tip contour," the midpoints of the left and right ends of the nose tip are selected; and for "nose wing width," the left and right wing points on the outer sides of the nostrils are selected.

[0032] Through this step-by-step mapping from aesthetic features to a set of geometric vertices, the system successfully acquired a structured and quantifiable adjustment dimension system. This lays the data foundation for all subsequent parameter-based precise control, transforming the stylistic adjustment of facial shapes from relying on the artist's subjective application to a systematic and procedural mathematical transformation of the spatial coordinates of these predefined vertex sets.

[0033] S102 generates a set of constraint parameters whose range of values ​​is controlled by using vertex coordinate offset as the sole quantization standard and setting offset thresholds that match aesthetic boundaries for the displacement of each vertex set on the X, Y, and Z axes.

[0034] Specifically, after completing the componentization and vertex mapping of facial features, the system enters the stage of generating constraint parameter sets. A precise and controlled quantization standard is established for each identified adjustment dimension. The system uses vertex coordinate offset as the unique quantization benchmark for all parameters. Following the nose example in step S101, the system defines the quantization standard for the nose as follows: for the "nose bridge shape" dimension, quantization is performed using the coordinate offset of the key vertex on the Y-axis (vertical direction), with positive values ​​indicating increased nose bridge height; for the "nose tip contour" and "nose wing width" dimensions, quantization is performed using the coordinate offset of the key vertex on the X-axis (horizontal direction), with negative values ​​indicating nose tip narrowing and nostril retraction, respectively.

[0035] To ensure strict adherence to aesthetic style, the system sets a closed interval for the displacement of each vertex set along each coordinate axis, defined by a minimum and a maximum offset threshold. These thresholds are validated through aesthetic modeling and visual verification of the selected style. For example, to ensure the nose is "high, delicate, and harmonious," the system sets a uniform range of [0.2, 0.9] for the Y-axis offset of key points in the bridge of the nose. An offset below 0.2 would cause the bridge of the nose to collapse, failing to meet the "high" characteristic, while an offset above 0.9 would appear exaggerated and distorted, deviating from the harmonious aesthetic of "trendy and cool." Similarly, corresponding thresholds are set for the X-axis offset of the tip and alar of the nose to ensure that any adjustments do not deviate from the visual boundaries of the target style. The system iterates through all predefined components and their vertex sets, applying corresponding threshold constraints to each set of coordinate offsets. Finally, all boundary-limited parameters are encapsulated into a structured and complete set of constraint parameters, constructing a style protection framework at the data level.

[0036] As an optional implementation of this application, optionally, in step S200, a preset face-shaping template is formed by assigning a fixed combination of parameters to a specific style and binding it with a preset association with facial decorations, based on the constraint parameter set, including: S201, based on the generated set of constraint parameters, forms the core facial feature preset for a specific style by assigning a fixed combination of parameters to that style.

[0037] Specifically, based on the constraint parameter set generated in step S102, which contains all facial shape parameters bounded by boundaries, the system initiates the construction process of the preset style template, transforming the abstract style concept into a concrete, executable data entity. The system first creates a style configuration template for each specific style defined in the plan (e.g., "cool and edgy short-haired girl style"). Then, for each facial component (e.g., nose, eyes, mouth) required to be represented by that style, the system assigns a qualified and fixed combination of parameters from the global constraint parameter set.

[0038] For example, such as Figure 2As shown, for the "nose" component, the system selects parameter ID 114 and assigns it a set of aesthetically validated, fixed vertex coordinate offset values; for the "eyes" component, it may select parameter ID 205 and assign it another set of fixed values. These selected parameter IDs and their corresponding fixed values ​​together constitute a unique identifier for the facial features of this specific style. This process ensures that all facial images generated by this style have highly reproducible and consistent geometric shapes of their core facial features, strictly adhering to preset aesthetic standards, and fundamentally avoiding potential morphological deviations when different users apply the same preset.

[0039] S202, by logically associating and packaging the core facial feature presets with the facial decoration presets specified in the system resource library at the data layer, a preset face-shaping template is obtained.

[0040] Specifically, after establishing the core facial feature parameter combination for a specific style through step S201, the system continues to advance the complete construction of the preset face-shaping template. To achieve uniformity and integrity in style expression, and going beyond simple geometric shaping, the system initiates a resource binding and data encapsulation process. This process operates at the system data layer, logically associating and packaging the determined core facial feature presets (i.e., fixed parameter combinations) with pre-stored facial decoration presets in the resource library that strictly match their visual tone. These facial decoration presets encompass non-geometric visual elements such as eyelash models, pupil textures, facial makeup textures, and hairstyle models. Through its internal data structure, the system establishes an inseparable reference relationship between the core facial feature parameters and these decoration resources.

[0041] For example, such as Figure 2 As shown, the facial features of a "cool and edgy short-haired girl" style are bound to an eyelash model named "NV12_EYELASH" and a makeup texture named "NV12_MAKEUP". Through this strong binding, a style preset that originally only contained geometric deformation data is materialized into a ready-to-use preset face-shaping template data package containing all necessary rendering instructions and resource references. When a user selects this template, the system can automatically load and apply all the geometric parameters and decorative resources it contains, providing the user with a highly coordinated and unified complete image experience with just one click, greatly improving the completeness of the user experience and operational efficiency.

[0042] As an optional implementation of this application, optionally, in step S300, based on multiple preset parts selected by the user, parameters are automatically decomposed and fused according to weight conflict resolution rules to generate a range of mixed style parameter values, achieving style diversity and controllability within a unified aesthetic framework, including: S301, based on the user's selection of multiple preset operation commands for various parts, automatically disassembles the specific feature parameters corresponding to these presets through the system, and obtains multiple sets of parameter data from different sources.

[0043] Specifically, the system interface layer receives and responds to the user's mixed selection operation command in the preset template library for multiple different parts (such as selecting the nose of preset A, the eyes of preset B, and the mouth of preset C). This command triggers the system backend to start an automated style fusion preprocessing process. The system first performs in-depth data analysis and decomposition on each preset part selected by the user. Based on the unique identifier of each preset, it traces back to its underlying data structure and extracts the specific feature parameters associated with that preset. These feature parameters are the quantized data represented by vertex coordinate offsets as defined in step S102.

[0044] For example, such as Figure 2 As shown, for the selected "Nose Preset NV12", the system parses its corresponding parameter ID114 and the specific offset values ​​of all vertices under that ID; for the selected "Eye Preset CY34", it parses parameter ID205 and its offset values. By traversing all selected feature presets and performing this parsing operation, the system ultimately obtains multiple sets of parameter data from different sources. These data sets constitute the raw input for style fusion calculation. There may be numerical overlaps and conflicts between them (such as two presets defining the same set of vertices but with different values), preparing the necessary data conditions for subsequent intelligent fusion calculation.

[0045] S302 uses preset weight conflict resolution rules to perform fusion calculations on the mixed parameter data and generate the value range of the mixed style parameters.

[0046] Specifically, after the decomposition and acquisition of multiple preset parameters are completed in step S301, the system enters the core style fusion calculation stage. The core task of this stage is to resolve conflicts between parameter data from different presets and generate a new, coordinated, and controlled parameter output range. The system relies on predefined weighted conflict resolution rules as the decision logic engine for the entire fusion process. This rule is a set of arbitration logic pre-set in the system to determine how to calculate the final valid value or value range when multiple presets provide different values ​​for the same parameter (i.e., the same set of vertices). After the fusion calculation process begins, the system traverses all conflicting parameter dimensions. For each conflict point, the system calls the weighted conflict resolution rules (e.g., adopting the "last-choose-first" principle, or performing a weighted average calculation on the conflicting values) to arbitrate and fuse the conflicting parameter values. The output of this calculation process is not a single, fixed parameter value, but a dynamically generated new parameter value range. This newly generated value range is a legal interval obtained after comprehensively considering the constraints of all input presets and being verified again by the global aesthetic framework. It inherits the controlled characteristics of the constraint parameter set.

[0047] Through this mechanism, the system ensures at the algorithm level that no matter how the user makes preset combinations, the parameters of all the facial images that may be generated will be strictly limited to the range of values ​​that conform to aesthetic standards after fusion. This gives users a high degree of customization freedom while reliably maintaining the style unity and visual harmony of the final output image.

[0048] Example 2 Based on the same principle as the aforementioned methods, a face image generation device based on a parametric style framework is also proposed, see [link to relevant documentation]. Figure 3 A facial image generation device 100 based on a parametric style framework according to an embodiment of this disclosure includes: The parameterized framework construction module 110 is used to obtain quantized constraint parameters by decomposing facial features through vertex coordinate offsets based on facial shape and aesthetic framework, and to generate a set of constraint parameters. The preset encapsulation and binding module 120 is used to form a preset face-shaping template by assigning a fixed combination of parameters to a specific style and binding it with a preset facial decoration according to the constraint parameter set. The style fusion calculation module 130 is used to automatically decompose parameters and perform fusion calculations based on multiple preset parts selected by the user, through weight conflict resolution rules, to generate a range of mixed style parameter values, thereby achieving style diversity and controllability within a unified aesthetic framework.

[0049] As an optional implementation of this application, the style fusion calculation module 130 may further include: The parameter decomposition and acquisition module 131 is used to automatically parse and extract the underlying parameter data based on multiple preset parts selected by the user. The weight allocation and conflict arbitration module 132 is used to perform weighted calculation and arbitration on conflicting parameters from different presets according to preset weight conflict resolution rules. The fusion range generation module 133 is used to calculate and output the new legal value range of each parameter under the fusion style based on the result of conflict arbitration.

[0050] Obviously, those skilled in the art should understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the control methods described above. The modules or steps of this application described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Therefore, this application is not limited to any specific hardware and software combination.

[0051] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the control methods described above. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium can also include combinations of the above types of memory.

[0052] Example 3 Furthermore, this application proposes an electronic device characterized in that it is used to implement any of the aforementioned facial image generation systems based on a parametric style framework, comprising: The processor is used to perform all computational tasks to implement a face generation system based on a parametric style framework; Memory is used to store processor-executable instructions and statically stored data.

[0053] The electronic device of this disclosure includes a processor and a memory for storing processor-executable instructions. The processor is configured to implement any of the preceding parametric style framework-based face generation systems when executing the executable instructions.

[0054] It should be noted that the number of processors can be one or more. Furthermore, the electronic device in this embodiment may also include input devices and output devices. The processor, memory, input devices, and output devices can be connected via a bus or other means, without specific limitations herein.

[0055] The memory, serving as a computer-readable storage medium for automated fault handling and self-learning methods in modules, can be used to store software programs, computer-executable programs, and various modules, such as the program or module corresponding to the face image generation system based on the parametric style framework in this disclosure. The processor executes various functional applications and data processing of the electronic device by running the software programs or modules stored in the memory.

[0056] Input devices can be used to receive input digital numbers or signals. These signals can be key signals related to user settings and function control of the device / terminal / server. Output devices can include display devices such as screens.

[0057] The various embodiments of this application have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for generating facial images based on a parametric style framework, characterized in that, include: Based on the facial shape and aesthetic framework, facial features are decomposed by vertex coordinate offset to obtain quantified constraint parameters and generate a constraint parameter set; Based on the set of constraint parameters, a preset face-shaping template is formed by assigning a fixed combination of parameters to a specific style and binding it with a preset association with facial decorations. Based on multiple preset body parts selected by the user, parameters are automatically decomposed and fused using weight conflict resolution rules to generate a range of mixed style parameter values. This achieves style diversity and controllability within a unified aesthetic framework, including: Based on the user's selection of multiple preset operation commands for various parts, the system automatically decomposes the specific feature parameters corresponding to these presets and obtains multiple sets of parameter data from different sources; By using preset weight conflict resolution rules, the mixed parameter data is fused and calculated to generate the value range of the mixed style parameters; By establishing predefined weight conflict resolution rules, the logical basis for the system to make automatic decisions is established; By applying the aforementioned weight conflict resolution rules to traverse and merge all conflict parameters, a range of mixed-style parameter values ​​that is subject to strict mathematical constraints is generated.

2. The facial image generation method based on a parametric style framework as described in claim 1, characterized in that, The facial shape and aesthetic framework decomposes facial features using vertex coordinate offsets to obtain quantized constraint parameters, generating a constraint parameter set, including: Based on the preset aesthetic framework requirements, a series of quantifiable adjustment dimensions are obtained by decomposing facial features into components to be adjusted corresponding to the key vertex set on the basic 3D model. By using vertex coordinate offset as the sole quantification standard and setting offset thresholds that match aesthetic boundaries for the displacement of each vertex set on the X, Y, and Z axes, a set of constraint parameters with controlled ranges of all parameter values ​​is generated.

3. The facial image generation method based on a parametric style framework as described in claim 1, characterized in that, The step of forming a preset face-shaping template by assigning fixed parameter combinations to specific styles and binding them with preset facial decorations, based on the constraint parameter set, includes: Based on the generated set of constraint parameters, the core facial feature presets for a specific style are formed by assigning fixed parameter combinations to a specific style. By logically associating and packaging the core facial feature presets with the facial decoration presets specified in the system resource library at the data layer, a preset face-shaping template is obtained.

4. The facial image generation method based on a parametric style framework as described in claim 2, characterized in that, The method uses vertex coordinate offset as the sole quantization standard and sets offset thresholds matching aesthetic boundaries for the displacement of each vertex set on the X, Y, and Z axes, generating a set of constraint parameters whose value ranges are all controlled, including: By mapping the style feature dimension to the positive displacement of a specific set of key vertices in the axial coordinates of the basic 3D model, the quantification of facial features is achieved. By setting a closed interval for the displacement of each set of vertex coordinates, which is strictly limited by the minimum and maximum offset thresholds, a set of constraint parameters that prevents style deviation and is mathematically strictly controlled is generated.

5. The facial image generation method based on a parametric style framework as described in claim 3, characterized in that, The process involves logically associating and packaging the core facial feature presets with the facial decoration presets specified in the system resource library at the data layer to obtain preset face-shaping templates, including: By establishing an inseparable data relationship between the fixed parameter combination representing the shape of the facial features in the database and the facial decoration preset representing the appearance, the inherent binding logic between style elements is constructed. By encapsulating all these associated style elements into data entities that can be directly called, you can obtain preset face-shaping templates that are ready to use and have a highly coordinated and unified presentation effect.

6. A facial image generation device based on a parametric style framework, characterized in that, The device includes: The parameterized framework construction module is used to decompose facial features based on facial shape and aesthetic framework, obtain quantized constraint parameters by vertex coordinate offset, and generate a set of constraint parameters. The preset encapsulation and binding module is used to form a preset face-shaping template by assigning a fixed combination of parameters to a specific style and binding it with the preset facial decoration according to the constraint parameter set. The style fusion calculation module automatically decomposes and performs fusion calculations based on multiple preset parts selected by the user, using weight conflict resolution rules, to generate a range of mixed style parameter values. This achieves style diversity and controllability within a unified aesthetic framework, including: Based on the user's selection of multiple preset operation commands for various parts, the system automatically decomposes the specific feature parameters corresponding to these presets and obtains multiple sets of parameter data from different sources; By using preset weight conflict resolution rules, the mixed parameter data is fused and calculated to generate the value range of the mixed style parameters; By establishing predefined weight conflict resolution rules, the logical basis for the system to make automatic decisions is established; By applying the aforementioned weight conflict resolution rules to traverse and merge all conflict parameters, a range of mixed-style parameter values ​​that is subject to strict mathematical constraints is generated.

7. The facial image generation device based on a parametric style framework according to claim 6, characterized in that, The style fusion calculation module also includes: The parameter decomposition and acquisition module is used to automatically parse and extract the underlying parameter data based on multiple preset parts selected by the user. The weighting and conflict arbitration module is used to perform weighted calculations and arbitrations on conflicting parameters from different presets according to preset weight conflict resolution rules. The fusion range generation module is used to calculate and output the new legal value range of each parameter under the fusion style based on the result of conflict arbitration.

8. An electronic device, characterized in that, For implementing the face image generation method based on a parametric style framework as described in any one of claims 1 to 5, including The processor is used to perform all computational tasks to implement a face generation system based on a parametric style framework; Memory is used to store processor-executable instructions and statically stored data.

Images