1. Introduction
Since the twenty-first century, improvements in quality of life have led to enriched and expanded spiritual and cultural needs. Visits to tourist attractions and cultural museums have skyrocketed, and interest in and exploration of culture has reached a new stage. Sanxingdui Culture, as an ancient culture found in Southwest China from the late Neolithic to the Shang and Zhou periods, has become the focus of academic and public attention because of its unique bronzes, gold and jade artifacts, as well as its mysterious figures, sacred trees, sacred beasts and other shapes. As an important place to display this cultural heritage (Reference ZhaoZhao, 2021), Sanxingdui Museum’s rich cultural relics resources provide valuable materials and inspiration for the design of cultural and creative products.
2. Overview of the Sanxingdui culture
Located in Guanghan City, Sichuan Province, the Sanxingdui site is one of the most important archaeological discoveries in China and the world. Samsangdui is famous for its stunning bronze, gold and jade artifacts that demonstrate the extraordinary craftsmanship and infinite creativity of the ancient Shu people (Reference GanGan, 2024). A large number of bronzes have been unearthed in Sanxingdui, such as bronze masks, bronze figures, bronze scared tree, etc. These bronzes have unique shapes and exquisite craftsmanship, showing the high development of bronze casting technology of ancient Shu civilization. Especially the bronze scared tree [Figure 1(a)], which is as high as 396 centimetres, is the largest single piece of bronze artifacts found in the world. Sanxingdui also unearthed gold wares such as gold sceptres and gold masks, as well as pottery, stoneware and ivory products, which provide valuable physical information for the study of ancient Shu civilization.
Figure 1. Sanxingdui culture survey. (a) The bronze scared tree1, (b) The bronze mask with protruding eyes
The ‘sacrificial pit’ of Sanxingdui has a unique sacrificial style in its excavated artifacts, and there are many priesthoods in the ritual activities of Sanxingdui of the ancient Shu civilization (Reference Guo, Xiang, Ran, Xie, Yang, Huang and DingGuo et al., 2023). The ‘protruding eyes’ are believed to be facial features of the Qiang people, and the bronze masks with protruding eyes [Figure 1(b)] are believed to be the physical manifestation of the ancestors of Ba Shu ‘Cancong’ is physically represented (Reference Li and WangLi & Wang, 2024). The discovery of the Sanxingdui site proves the existence of the ancient state of Shu three or four thousand years ago, reveals the unique Footnote face of the ancient Shu civilization, and shows that the Yangtze River basin, like the Yellow River basin, belongs to the mother of Chinese civilization, and is known as the ‘source of the Yangtze River civilization’, and the archaeological achievements of the Sanxingdui site provide a strong corroboration of the theory of the unity of the pluralism of the Chinese civilization (Reference Yang, Zeng and ChenYang et al., 2024).
3. AIGC technology applied in cultural and creative product design
The use of Artificial Intelligence has made significant advances in a wide range of design areas, transforming traditional approaches and providing new techniques to improve innovation and productivity (Li et al.). The core of AIGC technology lies in the use of machine learning algorithms, especially deep learning techniques such as Generative Adversarial Networks (GANs) and Variable Auto-Encoders (VAEs), to generate high-quality content. Through deep learning and image recognition technology, AIGC is able to integrate the unique genes of cultural symbols into modern design, realizing efficient and creative product design, and also helping to enhance the value of the products and user experience, promoting the sustainable development of the entire cultural and creative industry (Reference YuanYuan, 2024).
3.1. The development of AIGC technology
Chronologically, artificial intelligence began in 1950 when Alan Turing proposed that computers could simulate intelligent behaviour and think critically (Reference Bindra and JainBindra & Jain, 2024). In 1956, John McCarthy defined the term ‘artificial intelligence’ as ‘the science and engineering of making intelligent machines’ (Reference McCarthyMcCarthy, 1981) and developed the Lisp programming language, which had a significant impact on the design of the ALGOL programming language. Turing’s work became critical with the emergence of large-scale language models (LLMs), such as ChatGPT (GPT-4) or b Google Bard, that demonstrated the important ability to understand or coherently generate medical responses (Reference Segura, Segura, Porta, Heredia, Masquijo, Anain, Casola, Trevisson, Cafruni and L.Segura et al., 2024). Currently, AI technology has become a broad cross-disciplinary advanced science, which has been realized in manufacturing, education, security, finance, transportation, medicine and healthcare, gaming, film and television, entertainment and other technological fields.
Published in December 2016 by the Massachusetts Institute of Technology (MIT) Press, Deep Learning covers important topics such as neural networks, deep learning algorithms, and generative models (Reference GoodfellowGoodfellow, 2016). Midjourney Debuting in March 2022, Midjourney is an artificial intelligence program developed by the Midjourney Research Lab, and Midjourney’s current existence marks an important advancement in the role of AI technology in art creation, human-technology interaction, and social infrastructure Building Social Infrastructure (Reference Goodfellow, Pouget-Abadie, Mirza, Xu, Warde-Farley, Ozair, Courville and BengioGoodfellow et al., 2020).
3.2. AIGC technology in design
It wasn’t until 2018 that AI-generated content officially became a tool for art creation. A portrait titled Portrait of Edmond Bellamy (Reference RolezRolez, 2018), created by the French art collective Obvious using AI-generated content, caused a sensation in the art world when it sold at Christie’s for $432,500. The announcement of the Digital Art Prize winner, Space Opera House, a painting made from AI-generated images and retouched in Photoshop, accelerated the rise of AI-generated content in art creation (Reference Lyu, Wang, Lin and WuLyu et al., 2022).
In 2023, many globally renowned large-scale models represented by ChatGPT will be popular all over the world and begin to truly and comprehensively influence the cultural and creative industries. From passive influence to active action, most cultural enterprises begin to learn and set up generative AI-related projects to varying degrees, and the generative AI cultural and creative market is gradually determined (Reference Shaikh and MhaskeShaikh & Mhaske, 2023).
AIGC tech is crucial in element extraction from traditional culture, providing design materials and inspiration. It uses image recognition to identify patterns, motifs, and colours. It also aids in idea generation through machine learning, offering personalized solutions based on user preferences. AI-driven forecasting improves accuracy in predicting customer needs, ensuring smooth product manufacturing (Reference KoricanacKoricanac, 2021). During design, AIGC optimizes schemes, evaluates potential issues, and facilitates rapid iteration. It simulates lighting and usage scenarios to generate high-fidelity renderings, aiding timely design adjustments.
4. Methodology for the design of cultural and creative products
In the realm of genuine cultural and creative products, designers consistently map out the developmental trajectory of these items by adhering to the guidelines of aesthetics, tradition, and functionality (Reference ZhangZhang, 2021). With the continuous development of AIGC technology, the methodology of cultural and creative product design is undergoing a profound change. It can not only quickly extract and analyse cultural elements, but also automatically generate diversified design solutions, which greatly improves the efficiency of design and the richness of creativity.
4.1. General methods
Cultural and creative product design often involves imitation, pattern application, colour extraction, and citation. It begins with excavating and refining cultural elements, studying the target culture’s history, symbolism, and unique aspects (Reference LiLi, 2021). Representative elements are then extracted. Next, these elements are creatively conceptualized into a preliminary design. Market positioning follows, guiding subsequent steps. Sketches are drawn, materials selected, and a prototype created for feedback. Based on this, the design is optimized.
4.2. AIGC based design methodology
The combination of AIGC and cultural and creative product design has injected new vitality into the integration of traditional culture and creative industries. In the deep mining and intelligent extraction of cultural elements, the deep learning algorithm in AIGC technology is utilized to intelligently extract the cultural materials obtained from the research. Cultural elements with representativeness, uniqueness and universal recognition, such as patterns, colours and symbols, are extracted. Based on the extracted cultural elements, creative conceptualization is carried out to form a preliminary design concept. Then, using algorithms such as GAN and Image Generation in AIGC technology, multiple creative solutions are generated by inputting keywords or descriptions, which provide designers with more ideas and inspirations, and assist designers in their creative conceptualization. In the design sketch stage, AIGC technology is used to optimize the sketch to make it more in line with aesthetic and functional requirements. AIGC technology is used for material simulation and process optimization to improve the durability and environmental friendliness of the product. Finally, AIGC technology is utilized for product rendering to generate high-fidelity product renderings (Reference Ren, Qin and WangRen et al., 2023). By simulating different lighting conditions and usage scenarios, the design effect of the product is evaluated, and problems are identified and optimized in a timely manner.
5. Design practice
With the rise of cultural tourism and the booming development of cultural and creative industries, Sanxingdui Museum has achieved remarkable results in the development of cultural and creative products. However, how to create cultural and creative products that have cultural connotation and meet the market demand based on the inheritance of Sanxingdui culture, combined with modern aesthetic and technical means, has become an urgent problem to be solved. AIGC serves designers, designers and other departments use AIGC to collaborate, and use AIGC to collaborate in teams. The intention of this design is to propose a new general approach to the design of cultural and creative products, combining with AIGC to improve and optimize the process of product design. As AIGC platforms and models evolve, this common path can still be used.
In this research, the cultural elements of Sanxingdui were investigated and analysed. Subsequently, a Sanxingdui cultural product was designed through integration with the AIGC platform. A similar product with the highest sales volume from the Sanxingdui Museum was identified on an online shopping platform. The two products were then evaluated and compared using the Quality Function Deployment (QFD) method to determine which better aligns with customer needs [Figure 2(a)].
Figure 2. Design framework. (a) General design framework, (b) Sanxingdui cultural and creative product design workflow based on AIGC
5.1. Design practice of Sanxingdui cultural and creative products based on AIGC
This paper proposes a new way to design cultural and creative products by using AIGC. Firstly, elements are extracted according to the theme to be designed, and a database is sorted out. Secondly, instructions are input into the AI platform for it to learn. The designer verifies the AIGC scheme, iterates the design scheme according to the results, and finally uses AIGC to adjust the details [Figure 2(b)].
In the design of Sanxingdui cultural and creative products, the elements and textures of Sanxingdui are collected first. This design uses the dream AI to practice the design, and AIGC technology is first used in the creative conception stage. By inputting key words such as Sanxingdui culture theme, design style and use scene, AIGC model can automatically generate a variety of preliminary creative schemes [Figure 3(a)]. These solutions cover a variety of dimensions such as miniatures, colour matching, form design, etc. After the creative scheme is determined, the instructions are again input into the AIGC platform for intelligent optimization. The outputs can be evaluated the design details according to the designer’s optimization suggestions, such as increasing the plush texture of the product, highlighting traditional cultural connotations, and highlighting regional characteristics, etc [Figure 3(b)]. Once again selected the generated images in line with the theme, the model is more exquisite for detail repainting [Figure 3(c)]. This intelligent, targeted optimization of design can significantly improve design efficiency and shorten the cycle of repeated modifications.
Figure 3. AIGC generation procedure. (a) The creative, (b) Adjustment of design details, (c) Design details detail
5.2. Design results and validation
Quality Function Deployment (QFD) is an effective method for translating customer requirements into engineering characteristics. Quality Function Deployment is redefined as a methodology that converts consumer demands into ‘quality attributes’ and enhancing the design quality of the final product by systematically establishing the connections between these demands and attributes, beginning with the quality of each functional element and expanding this process to encompass the quality of each component and stage of production. The comprehensive quality of the product will emerge from this interconnected network of relationships (Reference AkaoAkao, 2024). QFD quality house can visualize the correlation between customer needs and engineering characteristics and the autocorrelation of engineering characteristics. At present, in the evaluation process of the relationship between customer needs and engineering characteristics, there are methods such as simple numerical scaling method, triangular fuzzy number or trapezoidal fuzzy number, fuzzy linear regression, etc., but in order to obtain the necessary data, it is necessary to invest a lot of time and cost.
The core tool of the traditional QFD method is the Quality House, which consists of the following parts. (1) Left wall: customer needs and customer needs importance. (2) Ceiling: product technical engineering characteristics. (3) Room: Relationship matrix and mapping from customer needs to product technical engineering features. (4) Roof: degree of autocorrelation between product technical engineering competitiveness. (5) Right wall: customer competitiveness assessment matrix. (6) Basement: technical competitiveness assessment matrix (Reference Taslim, Rendy and SusantoTaslim et al., 2021) (Figure 4).
Figure 4. House of quality
Based on the traditional QFD quality house, an improved quality house matrix is proposed in order to obtain the prioritization of product design features more conveniently and accurately, see Table 1.
Tabel 1. A modified HoQ matrix for design priority analysis
Where: CSNi represents customer satisfaction needs; Ci is the importance score of customer satisfaction needs. 1-2-3-4-5 are used to represent the five importance levels of extremely unimportant, unimportant, general, important, and extremely important, respectively. Where: PDQ represents product design characteristics; Pi,j represents the degree of correlation between customer satisfaction needs CSNi and product design characteristics PDQ. 0-1-3-5 are used to represent no correlation, weak correlation, medium correlation, and strong correlation, respectively. Representing the overall score of product design characteristics, the calculation formula is (Equation 1).
$$S = \Sigma \bar C_i P_{ij} $$
Customer Satisfaction Needs and Importance are the key inputs to the improved Quality House Matrix. Original descriptions of customer satisfaction need indicators for product styling are widely obtained through user and expert interviews. The experts remove and summarize the duplicated and similar indicators, and the indicators established after primary screening are analysed for importance using a 5-Likert scale method, which is completed by using questionnaire research and SPSS mathematical statistics analysis. The questionnaire adopts 5-Likert scale and SPSS mathematical statistics to obtain the mean and standard deviation of each index, to determine the importance of each index. In addition, the Cronbach α coefficient was used to test the reliability of the assessment results. If the final coefficient of each indicator is greater than 0.6, then it meets the reliability requirements, and the indicator and the total correlation coefficient is less than 0.7 of the demand for screening and finally combined with the average value of the importance of the indicator to obtain the importance of customer satisfaction demand rating.
Customer Satisfaction Needs Acquisition and Importance For the problem of cultural and creative product design, the original descriptions of 10 customer satisfaction needs indicators on the styling of cultural and creative products were collected through interviews, which are 1 good-looking, 2 cute, 3 fashionable, 4 attractive colour, 5 texture, 6 appropriate proportion, 7 personalization, 8 can see the elements of Sanxingdui, 9 embodies the culture of the Sichuan region, and 10 have a desire to buy. Experts will repeat and similar indicators for primary screening to establish 4 customer satisfaction needs: aesthetic needs 1, 4, 6; symbolic needs 8, 9; attractiveness needs 5, 10; style needs 2, 3, 7. 60 questionnaires were set up for the evaluation of the importance of customer satisfaction needs, and the questionnaires scored these 10 descriptions for the existing Sanxingdui Museum’s cultural and creative products (Figure 5(a)) and Sanxingdui’s creative and cultural products designed in this study (Figure 5(b)), respectively. Carry out the scoring of these 10 descriptions. SPSS software was used to calculate the mean, standard deviation and Cronbach’s α coefficient of the four customer satisfaction requirements.
Figure 5. Sanxingdui cultural and creative products. (a) Products from Sanxingdui Museum, (b) Products Generated by the AIGC Framework
For the products generated by AIGC, the results of the analysis are shown in Table 2.
Tabel 2. Mathematical and statistical results of AIGC-generated products
SPSS23.0 software was used to calculate the mean, standard deviation and Cronbach alpha coefficient of the four customer satisfaction requirements. The alpha coefficients of the first three indicators are greater than 0.7, which meets the requirement of reliability, and the Cronbach’s alpha coefficient of the fourth indicator is greater than 0.6 and less than 0.7, which is acceptable for reliability. The correlation coefficients of the four demands are greater than 0.4, so they do not need to be eliminated. The final customer satisfaction requirements are aesthetics requirement 11.13, style requirement 11.15, attractiveness requirement 11.22, and symbolization requirement 7.48.
For the official products of Sanxingdui Museum, the analysis results are shown in Table 3.
Tabel 3. Mathematical and statistical results of the products of the Samsung Mound Museum
SPSS23.0 software was used to calculate the mean, standard deviation and Cronbach’s α coefficient of the four customer satisfaction needs. The α coefficient of each index is more than 0.7, which meets the requirement of reliability, and the correlation coefficients of the four demands are more than 0.4, so they do not need to be eliminated. Finally obtain customer satisfaction needs: aesthetics requirements 9.28, style requirements 9.05, attractiveness requirements 6.05, symbolization requirements 6.20.
In this study, a systematic user demand acquisition system was constructed by integrating interview method, questionnaire survey and mathematical statistics analysis. Cronbach α coefficient (α>0.7) was used to verify the reliability of the data. This methodology design effectively guaranteed the scientific and reliability of the research results. According to the research data (Table 4), the overall score of design characteristics of cultural and creative products based on AIGC technology (31.45) was significantly higher than that of traditional Sanxingdui official products (22.72), and the difference was statistically significant (p<0.05). This finding may reveal the following important dimensions.
Tabel 4. Comparison table of overall product feature scores
Technological innovation dimension: AIGC algorithm generates a design scheme that not only maintains the bronze pattern characteristics but also integrates modern aesthetic elements through deep learning of the Sanxingdui cultural symbol system. This reconstruction of traditional elements may be more in line with Generation Z users’ consumption expectations of ‘traditional modernity’.
User involvement dimension: AIGC’s generative design allows for real-time adjustment of design parameters, making the product development process user accessible. According to the study, 82% of respondents believe that this ‘co-creation’ experience significantly increases the emotional value of the product.
It is worth noting that although the samples in this study meet the statistical requirements, there is no limit on the age structure of the respondents, which may affect the universality of the conclusion. Follow-up studies can be extended to users of all ages, and the accuracy of the subjective evaluation can be verified by objective measurement methods such as eye movement experiments. In addition, the long-term cultural transmission effects of AIGC products and their potential impact on traditional crafts remain to be observed.
6. Conclusions
During the research process, the uniqueness of Sanxingdui culture and the challenges of cultural and creative product design were analysed, alongside the potential application of AIGC technology in cultural and creative design. Through practical application, it was demonstrated that AIGC technology exhibits significant application value and potential in the design of Sanxingdui cultural and creative products.
Firstly, AIGC technology can significantly improve the efficiency of cultural and creative product design. Through automated generation of preliminary creative solutions, intelligent optimization of design details and other functions, AIGC technology greatly shortens the design cycle. Secondly, AIGC technology brings more diversified creativity to the design of Sanxingdui cultural and creative products. Thirdly, the application of AIGC technology in the design of Sanxingdui cultural and creative products also promotes cultural inheritance and innovation. By combining with modern design aesthetics, AIGC technology enables the cultural elements of Sanxingdui to be disseminated and inherited in a wider range, and at the same time inspires new creative sparks, which provides the impetus for the innovation of Sanxingdui culture.
But now the open AI platform is not perfectly enough to fully generate the effect that the designer requires, there are still technical aspects to be improved, due to the limitations of algorithm optimization, hardware support, etc., AI platforms would emerge errors when performing complex tasks (Reference Anantrasirichai and BullAnantrasirichai & Bull, 2022).
In conclusion, the practice of Sanxingdui cultural and creative product design based on AIGC provides new ideas and methods for the innovation and development of cultural and creative industries. Future research will continue to develop the framework, expand the types and fields of cultural and creative products, and contribute the practical wisdom and strength to cultural heritage and innovation.
Acknowledgments
This paper is sponsored by the project No. 24XJC760004, ‘Research on the Design of Sanxingdui Archaeological Image Database based on the AIGC Technology’, Research Project of Humanities and Social Sciences of the Ministry of Education of China. And the authors would like to express sincere thanks to Professor Haichao Li from Sichuan University for supporting the research work.
Open access
