Research on the Transformation and Innovation of Regional Art Characteristic Elements in Modern Digital Media Art

The core goal is to explore how to use modern digital technology to extract, transform, and innovate the characteristic elements of regional art so as to realize the deep integration of traditional art and digital media art [1, 2]. Regional art characteristic elements are not only the expression form of visual art but also unique symbols generated under a specific cultural background, bearing profound cultural connotations and historical value [3]. Its digital expression and innovative application can effectively promote the protection and dissemination of traditional art, provide rich sources of inspiration for digital art creation, and promote the further development of the cultural industry and the modern transformation of artistic creation [4, 5]. The technical route of the research is divided into two levels: theoretical analysis and technical realization. Based on semiotics, this paper analyzes the cultural semantics of regional art characteristic elements and systematically sorts out their historical origins, regional characteristics, and national aesthetics so as to construct the theoretical basis of digital expression [6, 7]. At the technical level, the research integrates advanced technical means such as deep learning, computer graphics, and human-computer interaction design, aiming at building a whole process system from artistic element extraction, digital storage, and intelligent generation to innovative expression [8]. This system transforms traditional artistic elements into digital language so that they can not only be preserved and reproduced but also be rejuvenated in the new creative environment, realizing the perfect combination of tradition and modernity [9].

The primary theoretical basis of the research includes pattern recognition, computer vision, and generative models. Pattern recognition technology provides technical support for automatic extraction of artistic elements; Computer vision realizes in-depth analysis of artistic elements through image processing and feature analysis; Generative models, incredibly generative adversarial networks (GANs), offer a wealth of technological possibilities for digital art creation [10, 11]. The combination of digital storage and blockchain technology provides a solution for the efficient management and secure storage of art data, ensuring the non-tampering and copyright protection of art resources. The coordinated development of these theories and technologies has laid a solid foundation for the realization of research objectives [12]. As an essential concept of the research subject, the Chinese element is based on the traditional cultural context and the necessary screening of Chinese traditional culture. From the foundation of tradition, the inheritance of modernity to the choice of cultural connotation, Chinese elements embody the unity of material culture and intangible culture [13, 14]. Through the analysis of the horizontal and vertical characteristics of Chinese elements, its core scope is preliminarily defined: horizontally includes visual art features such as composition, texture, and color, and vertically emphasizes its excellence and non-heterogeneity to ensure that it has universally recognized value as the core symbol of Chinese culture [15, 16]. Chinese elements also have the characteristics of "refinable symbolism, the positivity of national spirit and integrity of cultural implication." These characteristics provide an essential reference for the digital expression of regional art characteristic elements and modern art innovation so that they can not only be efficiently captured and expressed by digital technology but also display the unique charm of Chinese culture on the international digital art platform [17, 18].

2

Digital expression of characteristic elements of regional art

2.1

Definition of characteristic elements of regional art

In terms of technical realization, the definition of regional art characteristic elements depends on multi-modal data acquisition and analysis technology. Through high-resolution image acquisition, video recording and three-dimensional scanning technology, the visual characteristics of the elements are comprehensively recorded. As shown in equations (1) and (2), C_s,k are the normalized central distribution of feature weights, f_s is the feature function, V_i,k are the visual feature values, q is the smoothing parameter, V_i,j are the weighted integral of visual features, I(x,y) is the pixel value of the input image, and w_ij(x, y) are the weight distribution. Combined with natural language processing technology, the semantic annotation and analysis of art description documents are carried out to establish the corresponding relationship between cultural characteristics and visual characteristics.

(1)

V_{i, j} = \int_{0}^{L} (I (x, y) \cdot w_{i, j} (x, y)) d x d y

(2)

C_{s, k} = \sum_{i = 1}^{N} \frac{f_{s} (V_{i, k})}{□ V_{i, k} □ + q}

In texture acquisition, multi-angle acquisition and lighting condition change simulation technology can be used to ensure the comprehensiveness and accuracy of data; In color analysis, it is necessary to combine multispectral imaging and color mapping technology to extract the main color; As shown in equation (3), P_l,m are the multi-angle texture gradient direction F_x, F_y are the horizontal and vertical gradients, T_l,k are the texture feature values, and M is the number of samples. In composition analysis, the region segmentation algorithm can be used to analyze the composition elements of works of art in layers.

(3)

P_{l, m} = \sqrt{\frac{1}{M} \sum_{k = 1}^{M} (F_{x} T_{l, k}^{2} + F_{y} T_{l, k}^{2})}

Data processing and analysis require the application of advanced algorithms to further define element characteristics. Through the feature embedding method, as shown in equation (4), F_emb(x) is the feature embedding mapping function, W₁ and W₂ are the weight matrix, b₁ and b₂ are the bias terms, and m is the activation function, which can visualize multi-dimensional data, analyze the feature distribution among art elements in different regions, and then extract its internal law and uniqueness.

(4)

F_{e m b} (x) = W_{1} \cdot m (W_{2} \cdot x + b_{2}) + b_{1}

As shown in equations (5) and (6), D_total is the difference loss between data culture and actual output, E_cultural(t) is the theoretical characteristic, S(t) is the actual output characteristic, E_cultural is the interaction term between cultural characteristics and visual characteristics, s_m and q are weight parameters, m is the feature mapping function, A_m and N are input data, etiquette, spiritual connotation and values at the intangible cultural level; Vertically, it emphasizes its excellence, that is, cultural content with universal value, and non-heterogeneity, that is, content that can reflect the unique features of Chinese culture. This two-way definition ensures that Chinese elements are widely accepted without losing their unique national characteristics.

(5)

E_{cultural} = \sum_{m = 1}^{M} S_{m} \cdot t_{m} (A_{m}) + q \cdot m (N)

(6)

D_{total} = \int_{0}^{1} {(E_{cultural} (t) - S (t))}^{2} d t

2.2

Technical extraction methods of texture, color and composition in digital art

Texture extraction is an important means to analyze the texture and surface features of artistic works in digital art. As shown in equations (7) and (8), T(x,y) is the texture feature distribution, t_i is the feature weight, Gabor is the Gabor filter, a_i and t_i are the angle and wavelength, G_multi is the weighted combination of multimodal features, g_s is the feature function, V_s is the modal feature, Gabor filter and Laplace pyramid can effectively decompose the different scale and directional characteristics of texture. These methods capture details and overall features by simulating the human visual system's perception of texture.

(7)

G_{m u l t i} = \sum_{s = 1}^{S} g_{s} (V_{s}) \cdot g_{s}

(8)

T (x, y) = \sum_{i = 0}^{n} T_{i} \cdot Gabor (x, y, a_{i}, t_{i})

Fast Fourier transform technology can analyze the spatial frequency characteristics of texture, and provide quantitative description for the repeatability and directivity of texture. As shown in equations (9) and (10), L_lbp(x,y) is the local binary mode eigenvalue, I_p and I_c are pixel values, P is the number of sampling points, F_fft(u,v) is the fast Fourier transform frequency domain feature, f(x,y) is the spatial domain texture value, and M, N are image dimensions. Combined with the local binary mode and other methods, the subtle changes, granularity and directional features of texture can be accurately captured, providing support for subsequent classification and generation of artistic elements.

(9)

F_{F F T} (u, v) = \frac{1}{M N} \sum_{x = 0}^{M - 1} \sum_{y = 0}^{N - 1} f (x, y) e^{- j 2 π (\frac{u x}{M} + \frac{v y}{N})}

(10)

L_{L B P} (x, y) = \sum_{i = 0}^{P - 1} sgn (I_{p} - I_{c}) \cdot 2^{i}

Color extraction aims at analyzing the color relationship and emotional expression of artistic works. Through color space transformation technology, HSV and CIE-Lab model, color attributes can be better expressed from the perspective of human visual perception. As shown in Equation (11), C_hsv is an HSV color model, h, s and v are hue, saturation and lightness, and the HSV model can extract the perceptual characteristics of color by distinguishing hue, saturation and lightness; The CIE-Lab model more accurately reflects the physical characteristics of colors through brightness, chromaticity and contrast analysis.

(11)

C_{H S V} (h, s, v) = \frac{h \cdot v}{1 + \ln (1 + s)}

Composition analysis is the core of analyzing the spatial layout and structural aesthetic feeling of artistic works. As a classical algorithm, Hough transform can extract basic geometric shapes from works of art. As shown in equation (12), Ccie is the color difference of CIE-Lab model, and L*, a*, b*, L⁰, a⁰ and b⁰ are color space parameters, such as straight lines, arcs and axes of symmetry, thus revealing the basic structure of its composition. Combined with deep learning models and ResNet networks, the multi-level characteristics and artistic styles of complex compositions can be further analyzed.

(12)

C_{C I E} (L^{*}, a^{*}, b^{*}) = \sqrt{{(L^{*} - L_{0})}^{2} + {(a^{*} - a_{0})}^{2} + {(b^{*} - b_{0})}^{2}}

The deep learning model can automatically extract the artistic rules and style features hidden in the composition by learning a large number of composition examples. The analysis of the composition of Dunhuang murals can reveal its high symmetry and sense of hierarchy. As shown in equation (13), Hough(r,b) is the feature of Hough transformation, r, k, b is the parameter space, and P is the image pixel set, which provides reference for typesetting and layout in modern digital art creation. Image segmentation technology can also play a role in composition analysis. By separating the main composition elements in the picture separately, it is convenient for subsequent characteristic analysis and reconstruction.

(13)

Hough (r, b) = \sum_{(x, y) \in P} k (r - x \cos b - y \sin b)

3

The combination path of regional art characteristic elements and modern digital media art

3.1

Digital storage and management technology of regional art elements

The digital storage and management technology of regional art elements is the essential link to realize its transformation and innovation in modern digital media art, and it is also the core means to ensure the efficient utilization and security of data [19, 20]. With the increasing demand for the digitalization of regional art characteristic elements, large-scale and diversified art data puts forward higher requirements for storage and management. In this context, the combination of advanced distributed storage technology, semantic association annotation, and blockchain technology has become a critical path to achieving this goal [21, 22]. In terms of digital storage, distributed storage technologies, such as the Interplanetary File System (IPFS), provide efficient and reliable solutions by fragmenting data and replicating it on multiple nodes. This technology not only improves storage efficiency but also significantly enhances data disaster recovery ability and is suitable for the storage requirements of large-scale regional art element data [23, 24]. Figure 1 is an art-style generation diagram based on a generative adversarial network. In order to achieve cross-platform and cross-domain data sharing and interoperability, metadata standards (such as Dublin Core) are used to describe the stored digital art elements. Through this standard, the basic information of data, including source, creation background, and category, can be effectively standardized, and a unified semantic framework can be provided for subsequent retrieval and reuse [25, 26].

In the aspect of data management, the application of the semantic segmentation algorithm has dramatically improved the annotation accuracy of digital art elements. Algorithms such as DeepLab can accurately segment different parts of art images through deep learning models to distinguish specific elements such as textures, color blocks, and geometric shapes [27.28]. This segmentation method can provide high-quality input data for building a knowledge graph-based association retrieval system. Knowledge graph technology enables users to conduct multi-dimensional searches through keywords, context, and even semantic relationships by establishing the association relationship between art elements. Users can query related works based on the specific artistic style of a particular region and further explore their cultural background, artistic characteristics, and cross- regional relevance [29, 30]. For the security requirements of data storage and management, the introduction of blockchain technology provides unprecedented guarantees. Figure 2 shows the digital copyright protection of artworks supported by blockchain technology. Blockchain technology ensures that the stored content of digital art elements cannot be tampered with through asymmetric encryption and hash algorithms and, at the same time, provides support for the copyright protection of art elements. Each art data file can be stored on the blockchain and recorded in the form of timestamps and unique identifiers, thus ensuring the traceability of the source of the data and the authenticity of the content. This mechanism can also support value certification and income distribution in digital art transactions, and ensure the legitimate rights and interests of art creators.

Theoretical foundations cover semantic network theory, distributed storage models, and blockchain data management technologies. Semantic network theory provides a theoretical framework for the semantic association between fine art elements, enabling complex artistic content to be presented in a structured and intuitive way. The distributed storage model achieves the balance between storage efficiency and reliability through distributed deployment of nodes and data replication. Blockchain data management technology provides strong technical support for the secure storage and copyright protection of digital art elements through a decentralized trust mechanism. Chinese elements still face some challenges in the communication of digital art. The communication cultivation and stratification of the target audience are not perfect, the translation quality of digital artworks is low, there is an imbalance between economic value and cultural value, the development of digital art communication forms is uneven, and the degree of adaptability on Internet platforms is low. Table 1 is a comparison table of classification accuracy. These problems limit the potential of Chinese culture to achieve widespread dissemination through digital means to a certain extent. With the popularity of social media, short videos, live broadcasts, and mobile clients, these platforms provide unprecedented opportunities for the digital dissemination of regional art elements.

Table 1.

Comparison table of classification accuracy

Art Type	Paper-cut	New Year pictures	Wood grain
Linear function	76%	68.40%	45.60%
Polynomial function	76%	30.40%	45.60%
RBF function	76%	68.40%	45.60%
Sigmoid function	38%	30.40%	45.60%

3.2

Parametric modeling and visual presentation of digital regional elements

Parametric modeling and visual presentation of digital regional elements is a necessary technical means to integrate regional art characteristic elements into modern digital media art. Through digital expression and innovation of traditional art forms, they are endowed with new vitality in the digital world. Parametric modeling technology plays a central role in this process. It can accurately reproduce complex artistic textures, colors, and composition elements in digital form and realize flexible applications in different digital media. Parametric modeling firstly uses NURBS (non-uniform rational B-spline) and Bezier curve to model the complex shape of regional art elements accurately. The adjustable parameters of NURBS make it suitable for modeling traditional artworks with rich surface features, while Bezier curves are superior in local shape adjustment and smooth curve generation because of their simplicity and ease of use. Figure 3 is the digital storage efficiency evaluation diagram of regional art elements. These technologies can efficiently capture the geometric characteristics of regional art elements and lay the foundation for subsequent material and texture processing. In terms of material expression, PBR (physics-based rendering) technology plays an important role. It takes physical realism as the core, and through the simulation of the interaction between light and material, it restores the texture and details of traditional art, the smooth reflection of porcelain, the texture and touch of wood carving, etc., to the greatest extent.

In visual rendering, real-time rendering technology is the key. Taking the Unity and Unreal Engine engines as examples, these platforms provide powerful real-time rendering capabilities, which can dynamically display the light and shadow changes and material details of regional art elements. Ray Tracing technology further improves the rendering effect. By simulating the propagation and reflection of natural light, realistic effects such as dynamic light and shadow, natural shadow, and complex reflection are achieved. In order to meet the needs of immersive experience, virtual reality (VR) and augmented reality (AR) devices are introduced into the presentation process. Users can feel the details and spatial layout of regional art elements in an all-around way through VR devices, while AR technology can superimpose these art elements onto natural scenes to realize the dynamic combination of tradition and modernity. Figure 4 is the evaluation diagram of the disaster recovery capability of distributed storage nodes. The theoretical basis includes computer graphics, material modeling theory, and real-time rendering algorithms. Computer graphics provides algorithm support for modeling and rendering. Material modeling theory ensures the realism and consistency of artistic elements in the digital environment, while real-time rendering algorithm lays the foundation for real-time interaction and dynamic rendering.

In terms of communication strategy, the display of digital regional elements not only depends on technology but also needs systematic design around communicators, recipients, and content gatekeepers. Communicators need to uphold the "craftsman's heart" and improve their quality by carefully producing digital works of art. The recipient realizes more accurate cultural transmission through the "focus mode" so that different audiences can feel the artistic value and cultural connotation in the works. Through high-standard quality control, the content gatekeeper ensures that the works maintain their original artistic spirit and cultural essence in the process of delivery. The dissemination of digital art also needs to pay attention to the diversity of modern symbolic language expressions, use immersive experience technology to inherit the cultural context and adapt to users' needs for rapid content acquisition in the fragmented context. Combining the "communicable" cross-cultural communication language, the integrated communication mode under the background of media integration, and the Internet thinking that pays attention to users' sense of participation, a brand-new digital communication ecology with Chinese elements is constructed. Table 2 is a comparison table of training time. This ecology not only provides an operational path for the digital transformation and innovation of regional art characteristic elements but also contributes technical support and strategic methods for the global dissemination and exchange of Chinese traditional culture.

Table 2.

Training time comparison table

Art Type	Wood grain	New Year pictures	Paper-cut
One-Against-One	3.015 e-005s	1.459 e-005s	2.333 e-005s
One-Against-Rest	7.562 e-005s	5.092 e-005s	9.635 e-005s
DAG-SVMs	2.786 e-005s	1.976 e-005s	2.751 e-005s

4

Technical support in digital media art innovation

4.1

Art element generation method based on generative adversarial network (GAN)

Generative adversarial network (GAN) is a core technology in the field of current artistic element generation and innovation. It promotes the continuous optimization of the model by simulating the adversarial training between authentic images and generated images so as to achieve high-quality and creative artistic works. Generation. Under the background of modern digital media art, GAN can significantly promote the transformation and innovation of regional art characteristic elements, especially for the artistic creation of different cultures and styles. Through GAN, artists can automatically generate personalized artistic images with the help of deep learning technology, opening up new possibilities for artistic creation. As a mature generative adversarial network, StyleGAN can generate high-quality images with great artistic sense and visual impact. By controlling the layer level of the image, StyleGAN can finely adjust each element in the image, such as color, texture, form, etc., to produce stylized works of art. Figure 5 is an accuracy evaluation diagram of semantic segmentation and annotation of digital art elements. This feature is especially suitable for the extraction and transformation of regional art characteristic elements. It can combine traditional regional art styles with modern digital technology to create digital artworks that not only retain regional cultural characteristics but also have a modern sense of digital art.

CycleGAN has demonstrated its unique advantages in artistic style transfer. With CycleGAN, artistic styles can be converted between different images. Traditional regional art styles can be transferred to modern digital artworks so as to realize style integration and innovation under different cultural backgrounds. This style transfer is not limited to the transformation of tone or form but also includes the transfer of cultural symbols and artistic language, which enables the elements of regional art to cross the boundaries of time and space and integrate into the expression system of modern digital media art. Aiming at the unique style of regional art, emotional labeling and semantic segmentation technology are introduced into the research to enhance the cultural expressiveness of the generated images. Regional art styles are often closely related to specific historical and cultural backgrounds and emotional connotations. Therefore, accurately capturing and expressing these emotional and cultural elements is very important when generating artistic works. Through the introduction of emotional labels, GAN can take into account the dimension of emotional expression in the generation process so that the works not only visually present regional characteristics but also resonate with the audience at the emotional level. Figure 6 is a standardized evaluation diagram of art data in cross-platform sharing, while semantic segmentation technology can divide the image into local areas during the generation process so as to finely control the artistic effect of each element and ensure that the final image can fully reflect the rich connotation of regional culture.

4.2

Application of blockchain technology in copyright protection and value certification of works of art

The introduction of blockchain technology has opened up a new path for copyright protection and value certification of works of art. Through the technological advantages of decentralization, blockchain provides a highly secure and transparent solution for the digital storage, circulation, and trading of artistic works and solves the problems of information tampering and ownership ambiguity in traditional copyright protection methods. In terms of copyright protection, the application of Ethereum smart contracts enables each artwork to obtain a unique identification after digitization, and the ownership of the artwork is recorded on the blockchain in the form of non-homogeneous tokens (NFTs). As a carrier of digital assets, NFT can bind the ownership of artistic works to specific digital files, thereby ensuring the uniqueness and non-reproducibility of works in digital space. This method can not only prevent illegal copying and infringement of artworks but also provide creators with strong copyright protection through the non-tampering nature of on-chain data. The application of zero-knowledge proof technology (ZKP) enhances the privacy and security of transactions, allowing users to complete the purchase, transfer, or other commercial operations of artworks without exposing specific transaction information. In terms of data storage and verification, blockchain technology adopts the Merkle Tree structure to achieve efficient work storage and verification through a hierarchical hash algorithm. The transaction records of each work of art are added to the chain in the form of blocks, and its ownership and circulation process can be traced entirely. Figure 7 is a diagram of art element association retrieval and evaluation based on a knowledge graph. This technical design ensures the transparency and authenticity of ownership relationship and transaction information throughout the entire life cycle of artworks from creation to transaction. The structure based on Merkle Tree makes the data storage of large-scale works of art more efficient and can support rapid inspection and traceability, providing technical support for digital transactions of works of art.

The application of blockchain technology in art copyright protection and value certification is based on game theory, encryption algorithm, and digital copyright protection theory, which further promotes the deep integration of art and technology. The game theory model analyzes the behavioral strategies of different subjects in the blockchain platform to ensure the balance of the interests of artists, collectors, and consumers in copyright protection and value distribution. The encryption algorithm ensures the confidentiality and security of art data through asymmetric encryption technology. The theory of digital copyright protection supports regulations and standards for technology applications, enabling blockchain technology to comply with regulations in actual scenarios. The application of blockchain technology in the field of art also provides new possibilities for value certification. Through the art transaction history, ownership information, and authenticity certificate recorded on the chain, the value of works of art can be accurately evaluated. This method not only provides support for the transparency and standardization of the art market but also helps to reduce the risks caused by information asymmetry in traditional art transactions. Through the technical processing of art value certification, blockchain brings higher security and efficiency to the field of art investment and collection. Figure 8 is an evaluation diagram of application scenarios of blockchain technology in copyright protection. As the pillar of digital art, blockchain technology can also extend works of art from a single display function through accurate digital identity authentication and secure transaction processes. To a broader interactive experience, bringing a more realistic and immersive artistic perception to the audience. This combination of virtual and real communication form not only enriches artistic expression techniques but also provides a brand-new possibility for the blending of traditional and modern cultures.

5

Experimental analysis

In the experiment of texture, color, and composition extraction, the accuracy and efficiency of different extraction algorithms are mainly evaluated. Aiming at the extraction of image texture and color, the experiment quantifies the extraction accuracy of the algorithm by calculating the peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM). PSNR can measure the quality of the image, and the higher the PSNR, the lower the image loss and the higher the extraction accuracy; SSIM pays more attention to the structural similarity of images and can fully reflect the visual quality of images. In order to further evaluate the efficiency of the algorithm, the experiment also records the calculation time and resource occupation in the extraction process and analyzes the performance of different algorithms. Figure 9 is a storage node topology evaluation diagram of digital regional art elements. The experimental results show that the texture and color extraction algorithms based on deep learning are superior to traditional methods in accuracy and efficiency, especially in complex regional art element extraction, which can better retain the details and artistic features of images.

In the aspect of digital storage and management, the experiment verifies the stability of storage technology through query response time and data integrity tests. Query response time is an important index to measure the efficiency of the storage system. By querying a large number of artistic image data, the experiment tests the response speed of the system under high load. Figure 10 shows the image quality evaluation diagram generated in artistic style migration. The data integrity test focuses on whether the image data is lost or deformed during the storage process to ensure that the artistic work can maintain its original characteristics during the digital storage and management process. The experimental results show that the digital storage technology used has high stability and can effectively support the storage and management of large-scale artistic data.

6

Conclusion

This paper comprehensively probes into the transformation and innovation of regional art characteristic elements in modern digital media art and puts forward multiple paths of digital art creation through the combination of multiple technologies. By analyzing the digital expression, combination path, and technical support of regional art elements, this paper shows how digital technology can promote the application and innovation of regional art elements in artistic creation. The following are the main conclusions: 1)

The digital expression of regional art characteristic elements provides an essential way for the innovation and inheritance of traditional art forms. By extracting and analyzing the elements of texture, color, and composition of regional art, traditional art forms can be transformed into artistic content that adapts to the modern digital environment. The digital extraction method of texture and color can help artistic creators accurately reproduce regional styles and, on this basis, make artistic creation innovative. The digital processing of composition elements enables the visual language of regional art to be reconstructed and expressed on the digital platform. These technologies not only improve the efficiency of artistic creation but also enhance the communication power of regional art, enabling it to cross geographical and cultural boundaries and influence a wider audience.

2)

The combination of regional art elements and modern digital media art is the key to promoting the innovation and development of regional art. Digital storage and management technology can effectively solve the problems of preservation, access, and dissemination of artistic works and provide technical support for the broad application of regional art elements. On this basis, parametric modeling and visual presentation technology make regional art elements vividly present on the digital platform. Through digital technology, regional art elements can not only retain their original cultural connotation but also innovate and re-create in digital creation so as to achieve the organic integration of traditional and modern artistic styles. This process not only provides a new form of expression for regional art but also provides rich creative resources for modern digital media art.

3)

In the process of digital transformation of regional art characteristic elements, the accuracy of texture and color extraction is critical to artistic creation. By analyzing the running effects of different algorithms, the experimental data show that the average accuracy of the algorithm in texture extraction is 72.4%, while the accuracy of color extraction is 56%. In the extraction of composition elements, the structural similarity index (SSIM) of the image is 89.3, showing good visual quality. For the extracted artistic elements, the subjective evaluation results show that about 33% of the works meet the high artistic value standard. In terms of calculation time and resource occupancy, the average running time of the algorithm is 12.7 seconds, and the resource occupancy rate is about 47.8%. The artistic expressiveness of artistic images generated by generative adversarial networks (GAN) has been dramatically improved. The FID of the generated images IS 91, and the IS score IS 64.5, which shows that the generation effect IS better than the traditional method, and the artistic expressiveness IS obviously improved.

Idioma:: Inglés

Calendario de la edición:: 1 veces al año
Temas de la revista:: Ciencias de la vida, Ciencias de la vida, otros, Matemáticas, Matemáticas aplicadas, Matemáticas generales, Física, Física, otros

RSS Feed de revista

Research on the Transformation and Innovation of Regional Art Characteristic Elements in Modern Digital Media Art

Qian Fu

Publicado en línea: 27 feb 2025

Recibido: 08 oct 2024

Aceptado: 09 ene 2025

DOI: https://doi.org/10.2478/amns-2025-0119

Palabras claveRegional Fine Arts, Digital Media Arts, Generative adversarial networks, Blockchain technology

© 2025 Qian Fu, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Palabras clave
Regional Fine Arts, Digital Media Arts, Generative adversarial networks, Blockchain technology