Chinese traditional paper-cut art classification method based on multi-dimensional feature clustering analysis

Abstract

The invention discloses a Chinese traditional paper-cut art classification method based on multi-dimensional feature clustering analysis, and relates to the field of intangible cultural heritage digital protection. Paper-cut morphological features and calculation technique features are obtained through image processing, and expert scoring system quantification semantic features are constructed; k-prototypes clustering is improved by using a genetic algorithm, and a weight group is optimized and combined to design a hybrid distance metric function; constructing a form-content-value three-dimensional mutual construction model which comprises a form layer of technique complexity grading, a content layer of a cultural symbol association graph and a value layer of a design rule base; a feature database dynamic update interface is developed, a GAN (Generative Adversarial Network) is integrated to realize cultural and creative design, an online cultural semantic analysis engine is constructed, paper-cut multivariate feature quantitative analysis is realized, non-perpetual digital protection and cultural and creative innovation are promoted, and the method can be expanded to other fields.

Description

Technical Field

[0001] This invention relates to a method for classifying paper-cutting art, and more specifically, to a method for classifying traditional Chinese paper-cutting art based on multidimensional feature clustering analysis, belonging to the field of digital protection technology for intangible cultural heritage. Background Technology

[0002] The study of Chinese paper-cutting art has gradually formed a multi-dimensional and interdisciplinary academic framework, encompassing multiple fields such as digital preservation, intangible cultural heritage transmission pathways, regional artistic characteristics, interdisciplinary methodologies, and cultural and creative transformation. However, traditional paper-cutting art research has long relied on qualitative analysis methods, which have the following shortcomings:

[0003] The lack of quantitative correlation analysis of diverse characteristics such as form, technique, and cultural semantics makes it difficult to reveal the deep-seated laws governing the relationship between art form and cultural connotation.

[0004] In the development of cultural and creative products, the degree of automation in artistic feature extraction and semantic analysis is low, and the cost of manual intervention is high. Summary of the Invention

[0005] To address the aforementioned problems in existing technologies, this invention presents a classification method for traditional Chinese paper-cutting art based on multidimensional feature clustering analysis. The aim is to explore the relationships within paper-cutting art, revealing its classification rules and cultural connotations. This technology is expected to be extended to applications in digital preservation and innovative design in other fields.

[0006] To construct a three-dimensional mutual construction model of "form-content-value" to achieve the following goals: quantitative grading of paper-cutting techniques and correlation mapping of cultural semantics / structured storage and dynamic updating of intangible cultural heritage digital resources / providing feature library support for intelligent generation system of cultural and creative products.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solution:

[0008] 1. Multidimensional feature extraction

[0009] Morphological characteristics: Using image processing technology, the symmetry index (S) and the number of repeating units (N) of the paper-cut are accurately obtained. The symmetry index (S) reflects the degree of symmetry of the paper-cut pattern in space, and the number of repeating units (N) reflects the number of repeating patterns in the pattern. These indicators can describe the appearance characteristics of the paper-cut from a geometric perspective.

[0010] Technical characteristics: Calculate the proportion of positive engraving area (A), which is the proportion of the positive engraving part in the entire paper-cutting work area, to measure the degree of application of the positive engraving technique; calculate the complexity score (C) using the formula C = ΔS / Δt, where ΔS represents the area change and Δt represents the time change (in the paper-cutting creation simulation process). This score can reflect the complexity of the paper-cutting technique.

[0011] Semantic features: Construct a scientific and reasonable expert scoring system, invite experts in relevant fields to quantify the intensity of cultural connotations contained in paper-cutting works, with the score range being E∈[0,1], thereby transforming the abstract cultural semantics of paper-cutting works into calculable numerical features.

[0012] 2. Improved K-prototypes clustering

[0013] Traditional K-prototypes clustering algorithms have limitations when handling mixed data types. This invention optimizes the weight combination using a genetic algorithm. The genetic algorithm simulates biological evolution, searching for the optimal solution in the solution space of weight combinations through selection, crossover, and mutation operations, ensuring a reasonable weight allocation for different feature dimensions during clustering. Simultaneously, a hybrid distance metric function is designed to comprehensively consider the differences between numerical features (such as the numerical values ​​of morphological and technical features mentioned above) and categorical features (such as the cultural categories associated with semantic features) of paper-cutting, accurately measuring the distance between samples, thereby achieving effective clustering analysis of the multidimensional features of paper-cutting.

[0014] 3. Three-dimensional model construction

[0015] Formal layer: Based on the improved K-prototypes clustering output, the complexity of paper-cutting techniques is divided into four levels: G1-G4. G1 represents simple techniques, and G4 represents complex techniques. This hierarchical approach quantifies and standardizes the complexity of paper-cutting techniques.

[0016] Content Layer: Using semantic network analysis, a cultural symbol association graph is constructed, with cultural symbols in paper-cutting works as nodes and semantic relationships between symbols as edges. This graph can intuitively display the connections and hierarchical structure between different cultural symbols, and deeply explore the cultural connotations of paper-cutting works.

[0017] Value layer: Establish a design rule library, such as setting rules for matching G3 level paper-cutting with the theme of "harmony between man and nature", linking the skill level of paper-cutting with cultural themes and artistic value, and providing guidance for the inheritance and innovative application of paper-cutting art.

[0018] 4. Digital Applications

[0019] Development of a dynamic update interface for the feature database: Design and develop a dynamic update interface for the feature database to facilitate the timely and accurate entry of data into the database when new paper-cutting works or research results are acquired, ensuring the real-time nature and integrity of intangible cultural heritage digital resources, and realizing structured storage and dynamic updates.

[0020] Ensemble Generative Adversarial Networks (GANs) for Cultural and Creative Design: Driven by clustering results, Ensemble Generative Adversarial Networks (GANs) utilize the adversarial mechanism between the generator and discriminator in GANs to generate design schemes for cultural and creative products with paper-cutting art characteristics. The generator learns paper-cutting features and generates new designs, while the discriminator judges the similarity between the generated designs and the features of real paper-cutting. Through continuous iterative optimization, intelligent generation of cultural and creative products is achieved.

[0021] Building an online cultural semantic analysis engine: This engine allows users to access clustering analysis reports of paper-cutting works simply by scanning QR codes. Based on a pre-built semantic model and feature database, the engine quickly and accurately analyzes the cultural semantics of paper-cutting works, providing a convenient way for the public to understand the cultural connotations of paper-cutting.

[0022] Beneficial Effects: This study achieves quantitative analysis of the diverse characteristics of paper-cutting art, compensating for the shortcomings of traditional research methods and revealing the profound laws governing the form and cultural connotations of paper-cutting art. It improves the automation level of artistic feature extraction and semantic analysis in the development of cultural and creative products, reduces the cost of manual intervention, and promotes the innovative application of paper-cutting art in the cultural and creative fields. The constructed three-dimensional interactive model and digital application system provide a complete and effective solution for the digital protection of intangible cultural heritage, which is conducive to the inheritance and development of paper-cutting art. Furthermore, this technological approach can be extended to the protection and innovative design of other intangible cultural heritage fields. Attached Figure Description

[0023] Figure 1 This is the system architecture diagram of the present invention.

[0024] Figure 2 This is a three-dimensional inter-structure model diagram of the present invention.

[0025] Figure 1 The system architecture diagram illustrates the overall process from feature extraction, through clustering, model building to application interfaces. It includes modules for morphological feature, technique feature, and semantic feature extraction; modules for hybrid distance metrics, genetic algorithm optimization, and 3D hierarchical mapping involved in clustering; and application interface modules such as model building, feature database, GAN generator, and semantic parsing engine.

[0026] Figure 2 The three-dimensional interactive model diagram presents the relationship between the cultural symbol map, the design rule base, and the technique complexity. The cultural symbol map consists of elements such as symbol nodes, semantic edges, and network density, forming the content layer; the design rule base consists of elements such as the level-theme mapping table, the technique-symbol matrix, and the value assessment model, forming the value layer; and the technique complexity is divided into simple structure, basic composite, multi-layer nesting, and dynamic openwork, forming the form layer. The three are interconnected to construct the three-dimensional model. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] like Figure 1-2 The illustration shows a specific embodiment of a classification method for traditional Chinese paper-cutting art based on multidimensional feature clustering analysis. This embodiment of a classification method for traditional Chinese paper-cutting art based on multidimensional feature clustering analysis includes constructing a three-dimensional mutual construction model of "form-content-value" to realize the quantitative classification of paper-cutting techniques and the correlation mapping of cultural semantics, the structured storage and dynamic updating of intangible cultural heritage digital resources, and the feature library support of the intelligent generation system for cultural and creative products.

[0029] The method includes the following steps:

[0030] Step 1) Multidimensional feature extraction: Extract paper-cutting art features from three dimensions, which includes the following sub-steps:

[0031] 1.1) Morphological feature extraction: The symmetry index S and the number of repeating units N of the paper-cut are obtained through image processing, which are used to characterize the geometric features of the paper-cut appearance;

[0032] 1.2) Technique Feature Extraction: Calculate the area ratio A of positive engraving in paper cutting and the complexity score C, which are used to reflect the characteristics of paper cutting creation techniques; the calculation formula is C=ΔS / Δt;

[0033] 1.3) Semantic Feature Extraction: Construct an expert scoring system to quantify the cultural connotation intensity E of paper-cutting, E∈[0,1].

[0034] The cultural connotations of paper-cutting are quantified using the cultural symbolism E.

[0035] Step 2) Improve K-prototypes clustering: Optimize weight combinations using a genetic algorithm, design a hybrid distance metric function, and perform cluster analysis on the extracted multidimensional features;

[0036] Mixed distance metric functions:

[0037] Constraints: α+β+γ=1; α, β, γ∈[0,1];

[0038] Where Si and Si′ represent the standardized values ​​of continuous variables (such as the complexity of patterns in paper-cutting works), respectively, and similarity is quantified using Euclidean distance to reflect the morphological differences of traditional patterns; Cj and Cj′ represent the cultural attribute labels of paper-cutting (such as regional schools, themes (flowers and birds, auspicious patterns, etc.), respectively), and Hamming distance is used to determine category consistency; E k E′ k α, β, and γ represent the matching degree of binary variables, taking values ​​of 0 or 1, and measure the similarity of technology applications through the matching degree; α, β, and γ represent the weight coefficients optimized by the genetic algorithm, which are optimized by the genetic algorithm to adapt to different protection objectives (such as restoration priority or innovation priority).

[0039] In step 2), the following three methods are used to optimize the clustering process and perform clustering operations on the extracted features:

[0040] Hybrid distance metrics: Designing appropriate distance metrics to measure the degree of difference between feature data;

[0041] Genetic algorithm optimization: Utilizing genetic algorithms to optimize clustering-related parameters improves clustering performance;

[0042] 3D hierarchical mapping: Mapping feature data to a 3D space for hierarchical processing;

[0043] Step 3) 3D model construction: Based on the clustering results, a model is constructed to provide the infrastructure for subsequent applications. This includes the following sub-steps:

[0044] 3.1) Form layer construction: The complexity of paper-cutting techniques is classified according to the clustering output results, using G1-G4;

[0045] 3.2) Content layer construction: Constructing a relational graph of paper-cutting cultural symbols through semantic network analysis;

[0046] 3.3) Value layer construction: Establish a design rule library and associate technique levels with theme patterns.

[0047] Step 4) Digital application, which includes the following sub-steps:

[0048] 4.1) Feature Database Update: Design an application interface for dynamic updating of the feature database;

[0049] 4.2) Cultural and Creative Design Integration: Integrating Generative Adversarial Networks (GANs) to achieve clustering-driven cultural and creative design;

[0050] 4.3) Semantic parsing engine construction: Build an online cultural semantic parsing engine that supports scanning QR codes to obtain clustering analysis reports.

[0051] Step 4) The application interface contains three components:

[0052] Feature Database: Data interaction is achieved through the Application Programming Interface (API), used to store and manage paper-cutting art feature data;

[0053] GAN Generator: Employs the DCGAN architecture to generate content related to paper-cutting art (such as cultural and creative designs);

[0054] Semantic parsing engine: Using Natural Language Processing (NLP) services, it analyzes the cultural semantics of paper-cutting art.

[0055] In a preferred embodiment, the three-dimensional inter-structure model comprises three layers that are interconnected: the form layer, which is related to the complexity of the technique, and the complexity of the technique is divided into simple structure, basic composite, multi-layer nesting, and dynamic openwork; the content layer, which is composed of a cultural symbol map, which includes element symbol nodes, semantic edges, and network density; and the value layer, which relies on a design rule base, which includes a level-theme mapping table, a technique-symbol matrix, and a value assessment model.

[0056] Specific Example 1:

[0057] 1. Implementation of Multidimensional Feature Extraction

[0058] Morphological characteristics: Professional image processing software (such as a programming environment supported by the OpenCV library) is used to perform edge detection, contour extraction, and other operations on the digital image of the paper-cutting artwork. The symmetry index (S) is determined by calculating the number of axes of symmetry and the area ratio of symmetrical regions; pattern recognition algorithms are used to identify and count repeating pattern units to obtain the number of repeating units (N).

[0059] Technique characteristics: The paper-cut image is binarized to distinguish between positive and negative areas, and the area ratio of positive cuts is calculated by statistically analyzing the image pixels (A). During the simulated paper-cutting creation process, the area change ΔS and time change Δt at different creation stages are recorded, and the complexity score (C) is calculated according to the formula C = ΔS / Δt.

[0060] Semantic Features: A team of experts, including folklorists, art historians, and paper-cutting artists, was assembled to develop detailed cultural connotation scoring criteria, covering various aspects of paper-cutting, such as myths, folk customs, and auspicious meanings. Experts scored the cultural connotation intensity of paper-cutting works according to these criteria, resulting in a cultural connotation intensity value E.

[0061] 2. Improve the implementation of K-prototypes clustering

[0062] Genetic algorithm optimization of weight combinations: An initial population is set, with each individual representing a possible weight combination. The fitness of each individual is calculated, with the fitness function designed based on the compactness and separation of the clustering results, such as using metrics like the silhouette coefficient. The population is iteratively updated through selection (e.g., roulette wheel selection), crossover (e.g., single-point crossover), and mutation operations until the optimal weight combination is found.

[0063] Design a hybrid distance metric function: For numerical features (S, N, A, C), use Euclidean distance; for categorical features (the cultural category associated with E), use distance metrics suitable for categorical data, such as Minkowski distance. Combine the two distance metrics and construct a hybrid distance metric function through weighted summation to calculate the distance between samples.

[0064] 3. Implementation of 3D Model Construction

[0065] Formal layer: Based on the improved K-prototypes clustering results, paper-cutting works are classified according to the complexity of the technique, and divided into four levels: G1-G4. Detailed technique descriptions and evaluation criteria are formulated for each level.

[0066] Content Layer: A large number of paper-cutting works and related cultural background information are collected to construct a semantic network. Natural language processing techniques are used to preprocess the cultural description text, including word segmentation, part-of-speech tagging, and named entity recognition, to extract cultural symbols. The weights of semantic edges are determined by calculating the co-occurrence frequency and semantic similarity between symbols, thus constructing a cultural symbol association graph.

[0067] Value layer: Based on the level of paper-cutting techniques and cultural connotations, combined with market demand and artistic aesthetics, a design rule library is established. For example, through market research and expert consultation, it was determined that G3 level paper-cutting is suitable for combination with the theme of "harmony between man and nature" to develop related cultural and creative products.

[0068] 4. Implementation of Digital Applications

[0069] Development of a dynamic update interface for the feature database: A database management system (such as MySQL) is used to design the database table structure to store the multidimensional feature data of paper-cutting. A RESTful API interface is developed using web development technologies (such as Python's Django framework) to implement CRUD operations on the data, ensuring that new data can be entered into the database in a timely and accurate manner.

[0070] Enhancing Cultural and Creative Design with Generative Adversarial Networks (GANs): A mature deep learning framework (such as TensorFlow) is used to build a Generative Adversarial Network (GAN). Clustered paper-cutting feature data is used as input to the generator, which generates design sketches for cultural and creative products. A discriminator judges the similarity between the sketches and real paper-cutting features and the actual cultural and creative products. Through multiple iterations of training, the parameters of the generator and discriminator are optimized to obtain high-quality cultural and creative design schemes.

[0071] Build an online cultural semantic analysis engine: A semantic analysis engine is built based on natural language processing technologies (such as the BERT pre-trained model). Feature data of paper-cutting works and cultural description text are input into the engine, and the model inferences to obtain cultural semantic analysis results. Develop mobile or web application interfaces that allow users to scan a QR code to upload paper-cutting images and obtain clustering analysis reports and cultural semantic analysis content.

[0072] Finally, it should be noted that the present invention is not limited to the above embodiments, and many variations are possible. All variations that can be directly derived or conceived by those skilled in the art from the disclosure of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A classification method for traditional Chinese paper-cutting art based on multidimensional feature clustering analysis, characterized in that: Construct a three-dimensional mutual construction model of "form-content-value" to realize the quantitative grading of paper-cutting techniques and the correlation mapping of cultural semantics, the structured storage and dynamic updating of intangible cultural heritage digital resources, and the feature library support of the intelligent generation system for cultural and creative products. The method includes the following steps: Step 1) Multidimensional feature extraction: Extract paper-cutting art features from three dimensions, which includes the following sub-steps: 1.1) Morphological feature extraction: The symmetry index S and the number of repeating units N of the paper-cut are obtained through image processing, which are used to characterize the geometric features of the paper-cut appearance; 1.2) Technique feature extraction: Calculate the area ratio A of positive engraving in paper cutting and the complexity score C to reflect the characteristics of paper cutting creation techniques; 1.3) Semantic feature extraction: Construct an expert scoring system to quantify the cultural connotation of paper-cutting, E, where E∈[0,1], and use the cultural connotation E to quantify the cultural connotation of paper-cutting; Step 2) Improve K-prototypes clustering: Optimize weight combinations using a genetic algorithm, design a hybrid distance metric function, and perform cluster analysis on the extracted multidimensional features; Step 3) 3D model construction: Based on the clustering results, a model is constructed to provide the infrastructure for subsequent applications. This includes the following sub-steps: 3.1) Form layer construction: The complexity of paper-cutting techniques is classified according to the clustering output results, using G1-G4; 3.2) Content layer construction: Constructing a relational graph of paper-cutting cultural symbols through semantic network analysis; 3.3) Value layer construction: Establish a design rule library and associate technique levels with theme patterns. Step 4) Digital application, which includes the following sub-steps: 4.1) Feature Database Update: Design an application interface for dynamic updating of the feature database; 4.2) Cultural and Creative Design Integration: Integrating Generative Adversarial Networks (GANs) to achieve clustering-driven cultural and creative design; 4.3) Semantic parsing engine construction: Build an online cultural semantic parsing engine that supports scanning QR codes to obtain clustering analysis reports.

2. The classification method for traditional Chinese paper-cutting art based on multidimensional feature clustering analysis according to claim 1, characterized in that: Mixed distance metric functions: Constraints: α+β+γ=1; α, β, γ∈[0,1]; Where Si and Si′ represent the standardized values ​​of continuous variables, respectively, and similarity is quantified using Euclidean distance to reflect the morphological differences of traditional patterns; Cj and Cj′ represent the cultural attribute labels of paper-cutting, respectively, and Hamming distance is used to determine category consistency; E k E′ k α, β, and γ represent the matching degree of binary variables, taking values ​​of 0 or 1, and are used to measure the similarity of technology applications; α, β, and γ represent the weight coefficients optimized by the genetic algorithm, which are optimized by the genetic algorithm to adapt to different protection targets.

3. A classification method for traditional Chinese paper-cutting art based on multidimensional feature clustering analysis according to claim 1 or 2, characterized in that: In step 1), the formula for calculating C is C = ΔS / Δt.

4. The classification method for traditional Chinese paper-cutting art based on multidimensional feature clustering analysis according to claim 1, characterized in that: The three-dimensional mutual-structure model consists of three levels, which are interconnected. Form layer: related to the complexity of techniques, which is divided into simple structure, basic composite, multi-layer nesting, and dynamic openwork carving; Content layer: Composed of a cultural symbol graph, which includes element symbol nodes, semantic edges, and network density; Value layer: Based on the design rule base, which includes a level-topic mapping table, a technique-symbol matrix, and a value assessment model.

5. The classification method for traditional Chinese paper-cutting art based on multidimensional feature clustering analysis according to claim 1, characterized in that: In step 2), the following three methods are used to optimize the clustering process and perform clustering operations on the extracted features: Hybrid distance metrics: Designing appropriate distance metrics to measure the degree of difference between feature data; Genetic algorithm optimization: Utilizing genetic algorithms to optimize clustering-related parameters improves clustering performance; 3D hierarchical mapping: Mapping feature data to a 3D space for hierarchical processing.

6. A classification method for traditional Chinese paper-cutting art based on multidimensional feature clustering analysis according to claim 1 or 5, characterized in that: Step 4) The application interface contains three components: Feature Database: Data interaction is achieved through the Application Programming Interface (API), used to store and manage paper-cutting art feature data; GAN Generator: Employs the DCGAN architecture to generate content related to paper-cutting art; Semantic parsing engine: Using Natural Language Processing (NLP) services, it analyzes the cultural semantics of paper-cutting art.

Images