Virtual Exhibitions of Cultural Heritage: Research Landscape and Future Directions

Abstract

Virtual exhibitions of cultural heritage (CH) have become a key means for preservation, education, and global dissemination in the digital era. This study provides a comprehensive systematic review and bibliometric analysis of CH virtual exhibition research from 1999 to 2025. A total of 651 valid records were retrieved from the Web of Science Core Collection following the PRISMA 2020 guidelines. Three tools (CiteSpace, VOSviewer, and Bibliometrix) support stronger analysis. Results reveal that the field’s knowledge structure can be organized into the following three interrelated layers: (1) a technology-driven layer (laser scanning, photogrammetry, VR/AR, and multimodal interaction), (2) a systemic application layer (curatorial workflows, digital museums, and immersive storytelling), and (3) a user experience layer (educational impact, gamification, and trust building). These dimensions form a cyclical pyramid framework linking innovation, interpretation and perception. The study identifies persistent regional disparities, with China and Italy leading in publication volume, while countries such as Denmark and Australia achieve higher citation impacts due to advanced policy support and digital strategies. Emerging trends highlight the growing integration of gamified learning, AI-assisted curation, and immersive narrative design. These reflect a paradigm shift from technological demonstration to cultural interpretation. This study establishes a holistic analytical framework for understanding the evolution and future directions of CH virtual exhibitions, providing an essential reference for researchers, curators, and policymakers in the heritage informatics domain.

1. Introduction

Virtual exhibitions of cultural heritage (CH) combine with digital technologies to enable visitors to engage in exploration, educational activities, and interaction without physically accessing the heritage site. It provides tourists with a novel experience and offers researchers an effective medium, while also creating new opportunities for the dissemination and interpretation of CH values [1,2]. At present, mainstream technology-driven approaches to CH virtual exhibitions include virtual reality (VR), augmented reality (AR), additive manufacturing, and multi-sensory interaction supported by complex devices. These techniques allow CH institutions to design exhibitions entirely according to curatorial concepts and creative intentions, integrating a variety of additional content while being less constrained by the materiality or typology of CH objects [3,4]. Following the COVID-19 pandemic, the convenience and rapid cultural communication enabled by CH virtual exhibitions have further stimulated research and technological advancement in this field [5,6]. Numerous studies have shown that providing immersive virtual experiences significantly enhances visiting willingness of CH [7,8]. Currently, generate digital archives that have become one of the most essential tools for heritage preservation [9].

As a highly promising strategy for marketing and heritage conservation, virtual exhibitions of CH have evolved into labor-intensive endeavors involving multiple complex technologies. Firstly, CH encompasses a remarkably broad spectrum, including tangible elements (such as architecture, monuments, archival materials, artworks, and artifacts), intangible cultural heritage (ICH) such as folklore, traditional crafts, and knowledge systems, as well as natural heritage [10,11]. The diversity of CH forms makes it impossible to adopt a single technological approach when constructing digital archives. Secondly, variations in exhibition formats can greatly affect the effectiveness of presentation. For instance, VR technology is often well-suited for comprehensive visualization of historical buildings or communities, whereas certain forms of ICH may require additional interactive installations—such as pressure and infrared sensors—or multimedia equipment embedded in physical exhibition environments [12,13]. Finally, the cultural status of CH itself is the outcome of social processes; not every legacy from past generations qualifies as “heritage”. Its historical value is largely shaped by exhibition modes, dominant cultural narratives, and policy interventions [14,15]. Therefore, virtual exhibition technologies function not only as tools for marketing and preservation but also as transformative frameworks that reshape public perceptions of heritage, and further influence the very construction of its cultural value.

The above situation reveals the vast potential of CH virtual exhibitions for future development. However, most of the existing exhibitions remain at a limited conceptual stage or small-scale experiment. An effective model that can integrate technological, cultural, and policy factors has not been established. Although several literature reviews have emerged in this field, they have primarily focused on specific technological aspects, with insufficient attention paid to how these technologies interact with complex cultural and policy contexts. Moreover, the impact of new exhibition models on the interpretation of CH value has rarely been addressed.

For example, Wang et al. conducted an insightful analysis of gamification and pedagogical theories in cultural heritage; however, their work mainly summarized existing practical cases rather than constructing an integrated conceptual framework for the field. Similarly, Bekele et al. provided a comprehensive overview of VR and AR applications in CH, yet failed to establish a cohesive research structure, thereby revealing a significant gap that persists in the current body of literature [4,12].

Therefore, this study employs bibliometric methods and tools to answer the following questions:

• What developmental stages has the virtual exhibition of CH undergone to date, and what milestone events have marked its evolution?
• What roles have nations and policy frameworks, respectively, played in this process?
• How are the influences of journals and authors interrelated?
• What conceptual transformations are revealed through the evolution and clustering of keywords?
• What current research frontiers are indicated by the clustering of cited references?
• How might emerging exhibition models reshape the interpretation and articulation of CH values?

In addition, based on the results of the bibliometric analysis, this study further conducts an in-depth examination of the most actively explored research directions within the field and, on that basis, proposes forecasts for the future development trends of CH virtual exhibitions.

2. Method

This study follows the PRISMA 2020 guidelines [16] and consists of the following three main steps: data collection, data processing, and data analysis. The specific process of data acquisition and screening is illustrated in Figure 1.

For data collection, Web of Science Core Collection (WoS CC) was selected as the primary database. As one of the most widely used scholarly databases, WoS CC has been recognized by numerous researchers and institutions. In bibliometric studies, compared with other databases such as Scopus, WoS CC offers a unique citation data structure particularly well-suited for efficient knowledge mapping. In this study, efficient knowledge mapping refers to the systematic visualization and structural organization of bibliographic data that allows the rapid identification of research patterns, thematic relationships, and intellectual turning points across a large corpus (it emphasizes methodological efficiency in converting quantitative data into interpretable knowledge networks). Moreover, the introduction of features such as Keywords Plus enables the visualization of concept evolution through intelligent algorithms [17,18].

Accordingly, we designed an advanced search query within WoS CC, restricting the document types to “Article” and “Review” in order to ensure structural representativeness and to minimize data redundancy. After obtaining the initial search results, all records were manually screened based on titles and abstracts to exclude irrelevant content according to the following criteria:

• Publications entirely unrelated to CH were excluded.
• Studies related to CH but primarily focused on tourism or social media were also removed.

The manual screening process involved two main steps. First, due to minor overlaps in terminology, abbreviations, or occasional databases, some articles were identified as irrelevant to this study solely by reviewing titles. Such records were excluded directly from the search results. Second, we further examined the abstracts and full texts to exclude publications with ambiguous relevance. The database was accessed using the institutional account, and all data retrieval and downloads were completed on 14 September 2025, with the manual screening performed on the same day.

This process yielded 714 initial records, of which 651 remained after manual screening. The dataset spans the years 1999–2025 and includes 315 journals, 2213 authors, and a total of 30,179 cited references. All relevant metadata are summarized in

Table 1. Literature data details.

Data Types	Results/Values
Search formula	TS = (Online OR Virtual) AND TS = (Exhibition OR Show) AND TS = (Cultural heritage) AND DT = (Article OR Review) And LA = (English)
Direct search results	714
Post-filtering data	651
Sources	315
Authors	2213
Co-Authors per Doc	3.72
References	30,179

.

For data analysis, three software tools were employed: CiteSpace (6.4.R1), VOSviewer (1.6.20), and bibliometrix (5.1.1.9). Among them, CiteSpace, one of the most widely adopted tools in bibliometric research, has been extensively recognized for its capability to perform complex citation network analyses and generate clustering visualizations [19]. However, its performance in keyword analysis tends to be somewhat cumbersome. Therefore, the keyword clustering analysis was complemented using VOSviewer [20]. The remaining analyses—such as national collaboration networks, annual publication growth rates, and journal performance indicators—were conducted with bibliometrix to provide a more intuitive presentation of the results [21].

3. Results

3.1. Spatio-Temporal Analysis of Publications

3.1.1. Literature Time Span

According to the retrieval results, the earliest publication in this field appeared in 1999, and the annual growth rate has since maintained an average of 18.58%. Based on key achievements and milestone events within the temporal range, the development of CH virtual exhibitions is divided into three stages, aiming to establish an objective link between publication trends and historical events, and to provide a foundational understanding of the field’s evolution. Details are illustrated in Figure 2.

Phase 1 (1999–2013): This period marks the early stage of CH virtual exhibition development, during which foundational technologies gradually matured and conceptual frameworks were successively established. In 2003, the United Nations Educational, Scientific and Cultural Organization (UNESCO) adopted the Charter on the Preservation of Digital Heritage at its 32nd General Conference, signifying the global recognition of cultural heritage digitization [22]. Notably, the Charter emphasized the need to prioritize the protection of “born-digital” heritage—heritage created originally in digital form. Subsequently, the London Charter, released in 2006, became the first internationally acknowledged guideline for CH digitization. It aimed to ensure rigorous standards in computer-based visualization of CH and to reflect the unique properties of such technologies, while designating accessibility (to research, understand, interpret, preserve, and manage) as one of its guiding principles [23]. A milestone event occurred in 2013, when the International Conference on Digital Heritage—sponsored by UNESCO—was held in Marseille, France, marking a new era of global collaboration and academic recognition following more than a decade of groundwork [24].

Phase 2 (2014–2018): The establishment of standards and conceptual maturity triggered a technological revolution within the field. In 2014, Facebook announced its acquisition of Oculus VR, Inc. As the leader in VR technology, Oculus brought global attention and major investment to VR, even though its applications beyond gaming were still in the experimental stage [25]. In 2015, the British Museum hosted the world’s first “VR Weekend,” using Samsung VR headsets and tablets to recreate a Bronze Age village scene—representing the first genuine integration of museum collections with VR technology and the first practical implementation of CH virtual exhibition [26]. In 2016, Oculus released the first consumer-grade VR headset, Oculus Rift, capable of connecting to home PCs and offering a 90 Hz refresh rate. This breakthrough removed VR from the confines of museums and specialized institutions, significantly improving CH accessibility [27]. Subsequently, numerous similar practices emerged, such as Dreams of Dalí at the Dalí Museum in St. Petersburg, Florida [28], marking the beginning of widespread societal enthusiasm for CH virtual exhibitions.

Phase 3 (2019–2025): This phase represents the most dynamic and transformative period in the industry. In contrast to earlier emphasis on commercial potential, the devastating 2019 fire at Notre-Dame Cathedral in Paris underscored the critical importance of virtual technologies for CH preservation [29]. The outbreak of the COVID-19 pandemic further accelerated this shift: although nearly all museums and CH institutions expanded their online presence, the crisis exposed severe deficiencies in digital skills and infrastructure across the sector [30], thereby catalyzing rapid technological innovation and adoption. In 2021, Facebook rebranded as Meta, and the concept of the metaverse began merging with CH—a development widely regarded as the next phase of the internet [31]. Meanwhile, the rise in non-fungible tokens (NFTs) driven by blockchain technology has opened new avenues for CH exhibition and interpretation. Specifically, NFTs enable the certification of digital ownership and provenance for virtual artifacts, allowing museums and creators to authenticate, trade, and preserve cultural assets in decentralized environments. This mechanism not only expands exhibition formats but also redefines audience engagement through collectible and participatory digital heritage experiences, revealing the vast potential of this evolving field [32].

3.1.2. National Source of the Dataset

By analyzing the countries of origin for publications within the dataset using bibliometrix, it is possible to reveal the interconnections between geographical regions and policy contexts in this field. Specifically, the statistics of global publication volume and international collaboration were derived based on the corresponding authors’ affiliations, as illustrated in Figure 3. In the figure, darker color blocks represent higher publication volumes, while gray indicates countries with no recorded publications. The thickness of the red lines denotes the intensity of collaborative relationships between countries.

Table 2. Three indicators of the national literature.

PV	Country	TC	Country	AAC	Country
314	China	2149	Italy	62.20	Denmark
223	Italy	1036	China	39.50	Australia
141	Spain	809	Greece	35.00	Morocco
65	USA	791	Australia	32.30	Argentina
63	UK	718	Spain	30.00	Greece

lists the top five countries ranked by three key indicators: publication volume (PV), total citations (TC), and average article citations (AAC).

First, a pronounced imbalance can be observed across multiple indicators. Regionally, Europe, North America, and a few Asian countries occupy a dominant central position within the global publication and collaboration network, while most nations in South America and Africa exhibit very limited publication output—or even complete “data voids.” Second, the AAC, as a primary measure of scholarly impact, does not correspond symmetrically to publication volume. For instance, although China and Italy rank among the top two countries in terms of publication volume and total citations, neither appears in the top five for AAC.

This data structure highlights two key insights. First, it underscores the current deficiencies in technological capacity and knowledge sharing within the field. Bibliometric indicators cannot be treated as direct proxies for the quantity or significance of CH in a given region. Many developing or economically disadvantaged areas may possess CH of greater historical or cultural value but fail to attract international attention due to financial, technological, and sociocultural constraints [33,34]. Second, whether a larger number of publications necessarily equates to higher research quality remains an open question. For example, Australia and Morocco display exceptionally high AAC values, and Denmark records an outstandingly high figure (62.20), far above the mean. Yet, these countries do not stand out in either publication volume or international collaboration networks.

3.2. Source Journal and Author Data

Bradford’s Law describes how scientific literature in a discipline is unevenly distributed across journals, with a small number of core journals contributing most publications and citations [35,36]. According to the analysis by bibilimetrix, 22 journals within the scope of our research contributed five or more articles. The core sources zone shows the journals that have made significant contributions to the number of published articles (Figure 4). Based on TC and AAC, we further ranked the top five journals contributors in

Table 3. The top five journals ranked by TC and AAC.

PV	OA/N-OA	Source	Publisher	AAC	OA/N-OA	Source	Publisher
42	OA	Sustainability	MDPI	71	N-OA	Journal of Cultural Heritage	Elsevier
29	OA	Acm Journal on Computing and Cultural Heritage	ACM	31	OA	Acm Journal on Computing and Cultural Heritage	ACM
28	OA	Applied Sciences-Basel	MDPI	24	OA	Remote Sensing	MDPI
24	N-OA	Journal of Cultural Heritage	Elsevier	20	OA	Journal of Heritage Tourism	Taylor & Francis
22	OA	Heritage	MDPI	20	OA	Isprs International Journal of Geo-Information	MDPI

.

It can be observed that the journal influence reflected by Bradford’s Law is relatively limited. For instance, several journals identified as core according to Bradford’s Law and PV—such as Sustainability and Applied Sciences–Basel—do not appear among the top journals in terms of AAC. This indicates that these journals achieve high TC primarily due to their large PV, while their overall research impact remains comparatively low. Therefore, when considering all indicators comprehensively, journals such as the Journal of Cultural Heritage and the ACM Journal on Computing and Cultural Heritage should be recognized as the true core journals in this field.

Furthermore, the open access (OA) policies of journals play a crucial role in enhancing the democratization of scientific research [37,38]. Across all evaluated indicators, most of the core journals support full or partial OA. Notably, the publisher ACM has announced that the ACM Journal on Computing and Cultural Heritage will become fully OA by 2026 [39]. This development not only reflects OA’s emergence as a mainstream publishing model but also suggests that OA could serve as an effective means to alleviate the regional imbalances in research identified earlier.

3.3. Keywords Indicators

Using VOSviewer, we conducted a clustering analysis to construct the fundamental knowledge structure of the current research field (Figure 5). The full counting option was applied, allowing repeated keyword co-occurrences to be displayed and thereby enhancing the distinctiveness of the clusters. To ensure thematic focus and coherence, 139 keywords with a minimum frequency of five occurrences were selected from a total of 2938 keywords for inclusion in the analysis.

The results show that the keywords related to CH virtual exhibitions are evenly distributed and highly interconnected. In the visualization, the color and size of each node represent, respectively, the cluster to which the keyword belongs and the strength of its connections with other terms. Each time two co-cited articles share the same keyword, a link is established between them. To further clarify the knowledge structure,

Table 4. Keywords clustering.

Cluster ID	Theme	Total Number of Key Words	Representative Keywords
1	Cultural Heritage Tourism, Experience, and Management	40	heritage
			tourism
			management
			experience
			social media
2	Digitization and Virtual Reconstruction of Cultural Heritage	29	CH
			VR
			photogrammetry
			virtual reconstruction
			digitization
3	Informatization and Conservation Technologies for Cultural Heritage	29	model
			conservation
			artificial intelligence (AI)/deep learning (DL)
			semantic web/ontology
			intention/behavior
4	Augmented Reality and Virtual Museum Experience	26	AR
			museums
			virtual museum
			user experience
			cultural tourism
5	Virtual Reality Systems and Gamified Applications	10	VR
			system
			gamification
			serious games
			trust
6	Digital Cultural Heritage and Virtual Touring	5	digital cultural heritage
			virtual tour (VT)
			digital museum
			objects
			digital technology

summarizes all six clusters, including their thematic focus and representative keywords.

Overall, the clusters can be categorized into two primary dimensions: the application strategies and the practical technologies of CH virtual exhibitions.

At the application strategy level, research focuses on authenticity, perceived quality, and management aspects of virtual exhibitions, represented by Cluster 1 and Cluster 4. At the practical technology level, the structure is more complex: while both Cluster 2 and Cluster 3 demonstrate strong technological orientations, Cluster 2 emphasizes the presentation of digital content—featuring keywords such as VR and virtual reconstruction—whereas Cluster 3 concentrates on intelligent analysis and preservation strategies of CH itself, with representative keywords such as conservation, AI, and deep learning (DL). Clusters 5 and 6 establish a conceptual linkage between the technological and strategic domains, representing specific application pathways where technology actively drives exhibition strategies—illustrated by keywords such as gamification and virtual tour (VT).

In summary, the current keyword clustering reveals a three-tier structural framework within the research field, characterized by a mutually reinforcing relationship among the layers as follows:

• Bottom tier—Technology-driven: Represented by Clusters 2 and 3, this foundational level focuses on the development and innovation of immersive and digital technologies.
• Middle tier—Application scenarios: Represented by Clusters 1 and 4, this layer centers on concrete CH contexts and practices, including heritage tourism, museum exhibitions, and educational communication.
• Top tier—User experience: Extending across all clusters and most explicitly linked to Clusters 5 and 6, this layer addresses user experience and value perception in CH virtual exhibitions, where a complex evaluative framework is emerging.

3.4. Co-Cited Literature Clustering

A co-citation clustering analysis was conducted in CiteSpace to identify the most prominent research directions within the current dataset. Regarding specific parameters (K = 35) in the software. The result shows seven effective clusters. The reliability indices of the clustering are Q, S = 0.9182; typically, when this value is greater than or equal to 0.7, the clustering results are considered significant and robust. Additional detailed parameters are displayed in the upper-left corner of the visualization (see Figure 6).

Subsequently, the timeline view function in CiteSpace was employed to examine the temporal evolution of co-cited references within each cluster. The visualization of these temporal dynamics is presented in Figure 7.

In the visualization, the color bar at the lower left represents the temporal scale of citations, while each circle corresponds to an individual reference, with its size indicating the number of citations—the larger the circle, the higher the citation count. Accordingly, clusters that extend further to the right represent research areas with greater influence in recent years and thus denote topics of current scholarly interest. Conversely, clusters with shorter temporal spans suggest that their research content may have become outdated or has been absorbed into other clusters.

Based on this interpretation, Clusters 0, 2, and 7 remain highly active, with key references within them serving as intellectual foundations for the latest research directions. In contrast, Clusters 3, 4, 11, and 18 have received relatively few citations in recent years, suggesting that their topics have been assimilated into newer thematic clusters. Accordingly,

Table 5. Citation and cited literature for Cluster #0.

Citing Articles in Cluster #0		Cited References in Cluster #0
Author (Year)	Coverage	Author (Year)	Freq
Li, et al. [40] (2023)	30%	Lee, et al. [41] (2020)	16
Trunfio, et al. [42] (2023)	16%	Trunfio, et al. [43] (2022)	15
Li, et al. [44] (2024)	15%	Carvajal, et al. [45] (2020)	13
Lian and Xie [46] (2024)	13%	Ferdani, et al. [47] (2020)	9
Fissi, et al. [48] (2022)	12%	Errichiello, et al. [49] (2019)	9

, Table 6 and Table 7 list the top five cited and co-cited papers in Clusters 0, 2, and 7, followed by a detailed thematic analysis of each. In bibliometric terms, citing articles denote the publications in our dataset that reference other works, whereas cited references refer to the external sources being cited by these publications.

Cluster 0 primarily focuses on the digitization strategies of CH and the transformation of museums, with most publications in this cluster being review articles, representing the macro-strategic layer of CH virtual exhibitions. Current studies emphasize that the application of digital technologies—such as VR, AR, and various interactive devices—can significantly enhance audience engagement and CH accessibility. Particularly after the COVID-19 pandemic, the enormous potential of virtual exhibitions has become widely recognized among professionals in the field [50,51]. The emergence and high citation rates of multiple bibliometric studies indicate that the field is undergoing further conceptual and technological integration, striving to develop a comprehensive and panoramic understanding. However, research also highlights the complexity of virtual exhibition technologies and the financial and technical challenges encountered during museum transformation. Moreover, practitioners’ understanding of virtual exhibitions remains relatively limited [52,53]. In summary, virtual exhibition technologies have already exerted a profound impact on audience participation and the educational mission of CH and museums [54]. The key future research direction identified within this cluster lies in determining how to balance tradition and digital innovation, while further advancing the dissemination and preservation of CH.

Table 6. Citation and cited literature for Cluster #2.

Citing Articles in Cluster #2		Cited References in Cluster #2
Author (Year)	Coverage	Author (Year)	Freq
Jiang, et al. [55] (2024)	13%	Kim, et al. [56] (2020)	9
Yin, et al. [57] (2024)	11%	Correia Loureiro, et al. [58] (2020)	8
Jiang, et al. [59] (2025)	11%	Fan, et al. [60] (2022)	8
Wut and Ng [61] (2024)	10%	Atzeni, et al. [62] (2022)	7
Zainal Abidin, et al. [63] (2025)	9%	Hulusic, et al. [64] (2023)	6

Table 6. Citation and cited literature for Cluster #2.

Citing Articles in Cluster #2		Cited References in Cluster #2
Author (Year)	Coverage	Author (Year)	Freq
Jiang, et al. [55] (2024)	13%	Kim, et al. [56] (2020)	9
Yin, et al. [57] (2024)	11%	Correia Loureiro, et al. [58] (2020)	8
Jiang, et al. [59] (2025)	11%	Fan, et al. [60] (2022)	8
Wut and Ng [61] (2024)	10%	Atzeni, et al. [62] (2022)	7
Zainal Abidin, et al. [63] (2025)	9%	Hulusic, et al. [64] (2023)	6

Cluster 2 focuses on user behavior and acceptance mechanisms in virtual exhibitions, serving as a bridge between technological implementation and user experience. With the continuous integration of emerging technologies, accurately evaluating and consistently optimizing audience experience has become a persistent challenge. CH virtual exhibitions can be broadly categorized into two implementation types. On the one hand, digitized exhibitions are based on traditional museums, which are still anchored to specific CH sites but employ digital technologies to enhance visitors’ perception and provide highly immersive experiences [40,65]. On the other hand, fully virtual exhibitions or virtual tours are driven by various VR devices, aligning more closely with the concept of the metaverse, such as serious games or virtual classrooms. The key issue for this type lies in eliciting a sense of authenticity and strengthening cultural attachment [66,67].

Multiple studies have applied Structural Equation Modeling (SEM), fuzzy-set Qualitative Comparative Analysis (fsQCA), and the Stimulus–Organism–Response (SOR) framework to construct user feedback mechanisms, which have become mainstream analytical approaches in this domain [68,69,70]. However, these methodologies often face limitations in generalizability due to the strong regional specificity of CH, and their explanatory power remains insufficient for capturing the full complexity of user experience [71].

Overall, this cluster underscores the need to develop a more universal and efficient evaluation framework for user behavior and acceptance mechanisms in CH virtual exhibitions.

Table 7. Citation and cited literature for Cluster #7.

Citing Articles in Cluster #7		Cited References in Cluster #7
Author (Year)	Coverage	Author (Year)	Freq
Huang, et al. [72] (2025)	6%	Innocente, et al. [73] (2023)	12
Poggianti, et al. [74] (2025)	6%	Boboc, et al. [75] (2022)	8
Ge, et al. [76] (2025)	5%	Liu, et al. [77] (2022)	4
Yu, et al. [78] (2025)	5%	Page, et al. [79] (2021)	2
Hu and Ng [80] (2025)	4%	De Fino, et al. [81] (2022)	2

Table 7. Citation and cited literature for Cluster #7.

Citing Articles in Cluster #7		Cited References in Cluster #7
Author (Year)	Coverage	Author (Year)	Freq
Huang, et al. [72] (2025)	6%	Innocente, et al. [73] (2023)	12
Poggianti, et al. [74] (2025)	6%	Boboc, et al. [75] (2022)	8
Ge, et al. [76] (2025)	5%	Liu, et al. [77] (2022)	4
Yu, et al. [78] (2025)	5%	Page, et al. [79] (2021)	2
Hu and Ng [80] (2025)	4%	De Fino, et al. [81] (2022)	2

Articles in Cluster 7 are primarily case studies that focus on immersive system design and the educational functions of CH virtual exhibitions. Many of these studies showcase the most cutting-edge applications in the field, including immersive storytelling, gaze-based interaction, multi-user collaboration, and immersive experience workflows designed for people with disabilities. The industry as a whole is gradually establishing a comprehensive workflow that integrates digitization, curatorial design, immersive experience creation, and experience evaluation [82]. For instance, head-mounted displays (HMDs) often incorporate eye-tracking technology, while some VR systems can also capture users’ physiological signals, such as heart rate and body temperature. Compared with traditional questionnaire-based methods, this multi-sensor integration significantly enhances both the efficiency and reliability of user experience assessment, enabling real-time monitoring and optimization of virtual exhibition performance [83]. From an educational perspective, the lowering of technological barriers now allows user-centered content creation to reach a broader audience. As a result, the educational significance of CH can be greatly enhanced, further promoting digital democratization and sustainable development within the field.

3.5. Citation Burst Analysis

A citation burst analysis was conducted in CiteSpace to identify publications that experienced a sudden and significant increase in citation frequency within a specific time period. Such papers are typically published in leading journals of the field and often present groundbreaking insights or transformative conclusions that have substantially influenced the overall research trajectory. Figure 8 displays the citation burst chart generated by CiteSpace. In terms of software parameters (γ = 1.0, Minimum Duration = 2), were applied, resulting in the identification of five burst references. The detailed information and key findings of these publications are further summarized in

Table 8. Burst citation details.

Number	Author (Year)	Title	Key conclusion
1	Bekele, Pierdicca, Frontoni, Malinverni and Gain [12] (2018)	A Survey of Augmented, Virtual, and Mixed Reality for Cultural Heritage	One of the earliest comprehensive reviews in the field. Summarized the latest developments in AR and VR technologies. Identified the specific application domains of digital CH.
2	Jung and Tom Dieck [84] (2017)	Augmented reality, virtual reality and 3D printing for the co-creation of value for the visitor experience at cultural heritage places	Demonstrated that the integrated use of multiple technologies can significantly enhance co-creation value between CH institutions and audiences throughout the entire experience process. Highlighted the intense competition faced by traditional CH regions under digital transformation. Argued that traditional CH institutions should cultivate positive public reputation and reassess their institutional intent.
3	Carvajal, Morita and Bilmes [45] (2020)	Virtual museums. Captured reality and 3D modeling	Developed a virtual CH creation and experimental environment based on the Latin American CH context. Showed that photogrammetry-based 3D models can effectively reduce polygon counts while maintaining photorealistic quality. Established an efficient virtual experimental environment that enables rapid and multifunctional virtual exhibition design.
4	Trunfio, Lucia, Campana and Magnelli [43] (2022)	Innovating the cultural heritage museum service model through virtual reality and augmented reality: the effects on the overall visitor experience and satisfaction	Constructed the relationship between museum service innovation models and visitor satisfaction. Demonstrated the disruptive impact of AR and VR technologies on CH museum service paradigms. Initiated the academic discourse on “Museum 4.0.”
5	Bozzelli, et al. [85] (2019)	An integrated VR/AR framework for user-centric interactive experience of cultural heritage: The ArkaeVision project	Developed ArkaeVision, an integrated digital experience platform offering multiple modes of engaging with CH. Achieved integrated operation of CH virtual exhibitions and established a comprehensive, effective workflow.

.

4. Empirical Analysis and Results Interpretation

4.1. Analysis of Theme Evolution

Based on the temporal segmentation of publication periods, bibliometrix was employed to analyze the emerging, evolving, and declining themes within each stage. In order to clarify the conceptual transformations of the research field across different time spans, the results are illustrated in Figure 9.

The visualization adopts the same three-stage division described in Section 3.1.1. Each colored block represents the clustered keyword themes corresponding to a given time period. When research progresses to the next stage, arrows are drawn between clusters to indicate merging or thematic inheritance; the thicker the arrow, the stronger the continuity of keywords between the two clusters. Figure 10 provides a detailed representation of the clustered keywords across Phases 1, 2, and 3.

Overall, the thematic evolution of CH virtual exhibitions exhibits a pattern of initial expansion followed by convergence.

In the first phase (1999–2013), research themes were relatively broad, with most keywords centering on technological exploration and small-scale experimentation, such as laser scanning (LS), 3D modeling, and texture mapping. This indicates that studies during this period primarily focused on establishing digital archives of CH, largely driven by advancements in CH detection technologies. In addition, keywords such as “worldwide web” and “web application” suggest that early efforts were already made to disseminate CH values and information through the Internet. The launch of Google Arts and Culture in 2011 exemplified such early practices—it employed Google Street View technology in collaboration with museums worldwide to provide high-resolution artwork imagery and virtual gallery tours, demonstrating both the conceptual significance and vast market potential of CH virtual exhibitions [86,87].

During the second phase (2014–2018), the number of research themes increased sharply, forming 15 distinct clusters. Some keywords from the first phase evolved into independent thematic clusters. For example, laser scanning shifted its relational position, while photogrammetry—previously subordinate to that cluster—became highly frequent and formed a new cluster along with keywords such as “documentation” and “sites”. Meanwhile, information retrieval and web applications disappeared, suggesting that Internet applications had by then become a fundamental component of CH digitization. Under the impetus of emerging technologies, research displayed a pronounced interdisciplinary integration [88]. A representative initiative from this period is the Smithsonian Digital Volunteers program, a crowdsourced digital transcription platform enabling volunteers to transcribe historical archives, books, manuscripts, and reports to enhance the accessibility and searchability of CH in the digital era [89,90]. To date, the platform has attracted over 100,000 volunteers, who have transcribed more than 1.5 million pages of archival material [91]. The emergence of such CH crowdsourcing models vividly illustrates the broad application of digital technology and the thematic diversification that characterized this stage.

In the third phase (2019–2025), the total number of keywords rose dramatically (251), yet only three effective clusters were formed, indicating that current research directions have become more consolidated and stable. The largest cluster is labeled “heritage”, featuring high-frequency keywords such as “tourism”, “model”, “experience”, and “social media”. Notably, ICH, which did not appear in the first two phases, reached a frequency of 18 in this stage and clustered with terms like social media. This reflects how technological advancements have significantly enhanced the exhibition potential and cultural visibility of ICH, positioning it as one of the most prominent heritage categories for future research [92,93]. For instance, in a project centered on Hong Kong’s Hungry Ghosts Festival, researchers integrated the festival’s historical origins, rituals, stage performances, and environmental atmosphere within a virtual space, employing multimodal storytelling to achieve a participatory design of the ICH exhibition—an approach representative of the latest mainstream practices [94].

To clarify this evolutionary process,

Table 9. Top 10 keywords at different stages (the numbers in brackets indicate the frequency of each keyword’s occurrence).

1999–2013		2014–2018		2019–2025
Keywords (Occurrences)	Subject Category	Keywords (Occurrences)	Subject Category	Keywords (Occurrences)	Subject Category
cultural heritage (9)	CH	cultural heritage (18)	CH	heritage (27)	heritage
virtual reality (8)	CH	cultural heritage (8)	CH	tourism (26)	heritage
augmented reality (5)	CH	heritage (6)	CH	model (25)	heritage
virtual museum (4)	CH	virtual reality (6)	CH	experience (22)	heritage
cultural heritage (3)	CH	augmented reality (5)	CH	management (20)	heritage
museums (4)	museums	models (5)	CH	social media (20)	heritage
heritage (3)	museums	museums (5)	CH	ICH (18)	heritage
laser scanning (3)	LS	photogrammetry (5)	photogrammetry	conservation (15)
photogrammetry (3)	LS	art (4)	art	cultural tourism (15)	heritage
3d modeling (2)	LS	museum (3)	Art	impact (15)	heritage

lists the top ten most frequent keywords for each phase.

4.2. Analysis of Geographical Characteristics of Publications

Building upon the analysis in Section 3.1.2, we further explored the interconnections among national influence, CH resources, and policy frameworks. First, China and Italy stand out prominently in both PV and TC, as each possesses some of the world’s most extensive CH resources—Italy with 60 and China with 59 UNESCO-recognized World Heritage Sites. From a policy perspective, Italy’s establishment of the L’Istituto centrale per la digitalizzazione del patrimonio culturale—Digital Library in Rome in 2020, under the General Directorate for Digitization and Communication, represents a milestone initiative aimed at facilitating the digital transformation of CH institutions and improving preservation, management, and public engagement [95]. Unlike conventional CH or museum programs, this initiative emphasizes pragmatic, sustainable implementation, including assisting institutions in building digital portals, connecting them with digital service companies and app developers, and incorporating artificial intelligence (AI) into archival management. Such enduring and integrated policy measures have been key factors behind Italy’s leading position in this field [96].China, the top contributor in publication volume, benefits both from its rich domestic CH resources and from its top-down national policy structure. Most of China’s heritage-related policies are broad in scope and mandatorily enforced by local governments. For example, the 2022 policy document Opinions on Promoting the Implementation of the National Cultural Digitization Strategy established strict data security standards and required the creation of national digital archives and cloud service platforms across local governments and cultural institutions [97]. In contrast, Denmark, despite having only a dozen or so World Heritage Sites, ranks exceptionally high in AAC. While isolated, highly cited papers may partially exaggerate this index, Denmark’s leadership nonetheless stems from its advanced CH management philosophy and forward-looking policy continuity. For instance, the Royal Danish Library not only completed its 2020–2023 Material Cultural Heritage Digitization Strategy but also launched the 2023–2026 research program Cultural Heritage and Memory Technologies, which emphasizes the increasingly hybrid and fluid boundaries of CH in material media. It argues that preservation and translation practices may themselves create new heritage or redefine existing cultural meanings [98]. Such remarkable policy continuity and conceptual innovation explain Denmark’s “small but sophisticated” leadership position—shared by a few similarly progressive nations.

Overall, at the national and policy levels, the abundance of CH resources within the country, capital investment, the continuity and compulsion of policies, and the cutting-edge nature of CH management concepts are the four key factors determining the contribution of CH virtual exhibitions from this perspective. Empirical evidence shows that the majority of leading contributors are developed or upper-middle-income countries, revealing a pronounced global imbalance in financial investment and social participation—a continuing challenge for the field [99]. Furthermore, policy guidance does not always correlate directly with research impact. Some countries demonstrate high PV but lack corresponding advantages in TC or AAC. While not all research can be guaranteed to hold lasting value, this discrepancy highlights the critical importance of cross-sector and international collaboration. Isolated policy or financial support often fails to generate proportional scientific or commercial returns. Hence, these multiple driving forces are cyclical and mutually reinforcing. While some studies may emphasize CH preservation, tourism, consumption, or policy design, all ultimately contribute to the broader cultural heritage value cycle. In turn, advanced management philosophies foster the discovery and redefinition of new CH values [100].

4.3. Analysis of Keywords Clustering Results

As discussed earlier, a three-tier (technology-driven tier, application-scenario tier, and user-experience tier) conceptual framework was established based on the keyword clustering results. In this section, we provide a detailed analysis of the current research status and existing bottlenecks at each hierarchical level.

4.3.1. Bottom Tier: Technology-Driven

Based on the hierarchical framework outlined above, the bottom tier—the technology-driven research layer—can be further divided into the construction of digital archives for CH and implementation pathways of CH virtual exhibitions.

(1)

The construction of digital archives for CH

The establishment of digital archives forms the foundation of CH virtual exhibitions, encompassing the digitization, modeling, information storage, and archival management of various CH types [101,102]. Following UNESCO’s classification of CH,

Table 10. The establishment method of digital heritage archives.

CH Type	Acquisition Technology	Representative Devices	Resolution/Accuracy	Applicable Scenarios	Technical Difficulties
Building heritage	Terrestrial Laser Scanning (LiDAR); Close-range Photogrammetry	RIEGL VZ-400i (RIEGL Vienna, Austria)/Leica P50 TLS (Leica Geosystems AG Heerbrugg, Switzerland); UAV DJI M300 RTK (DJI Shenzhen, China) + Sony α7R camera (Sony Corporation Tokyo, Japan)	Point spacing ≤ 3 mm @ 10 m; model accuracy ±2–5 mm	External morphology of large-scale architecture Historical sites and monuments	Large data volume and high processing complexity Limited accuracy in multi-station point cloud registration
Archaeological heritage	Structured-light scanning; Ground-Penetrating Radar (GPR); Computed Tomography (CT)	Artec Eva (Artec 3D Luxembourg City, Luxembourg)/Creaform Go!SCAN 50 (Creaform Inc. Lévis, Canada); GSSI SIR-4000 GPR system (Geophysical Survey Systems Inc. Nashua, USA)	Spatial accuracy 0.1–0.5 mm (surface); depth resolution 10–20 cm (GPR)	Archaeological excavation layers and remains Small- to medium-sized artifacts Subsurface heritage detection	Challenging on-site environmental conditions Fragile artifacts; certain materials (e.g., metal, highly reflective surfaces) unsuitable for LiDAR
Painting and mural heritage	Multispectral Imaging (MSI); Reflectance Transformation Imaging (RTI); X-ray Fluorescence (XRF)	Specim FX10 MSI camera ((400–1000 nm) Specim, Spectral Imaging Ltd. Oulu, Finland); RTI hemispherical lighting array (Custom-built, following CHI RTI specifications Chicago, USA); Bruker TRACER XRF (Bruker Corporation Billerica, USA)	Spectral band width 5–10 nm; pixel resolution < 50 µm	Archival and museum collections Mural texture and pigment distribution Revealing underdrawings and restoration layers	Non-contact scanning required to prevent damage Difficulty in aligning MSI data with 3D models
ICH	Optical/Inertial Motion Capture; Volumetric Video/Spatial Audio Capture	Vicon V5 ((240 fps) Vicon Motion Systems Ltd. Oxford, UK); OptiTrack Prime 17W (NaturalPoint Inc. Corvallis, USA); Insta360 Pro 2 ((8 K 3D video) Arashi Vision Co., Ltd. Shenzhen, China)	Positional precision < 1 mm; temporal latency < 10 ms	Recording of traditional performance movements Documentation of craft-making processes Oral heritage and musical traditions	Intrusive interference of equipment with performers Insufficient spatial–temporal precision of captured data

summarizes the mainstream digital acquisition techniques, typical tools, applicable scenarios, and current challenges associated with different categories of heritage.

It can be observed that current methods for constructing digital archives of CH are characterized by a high degree of technological integration and cross-disciplinary application. For instance, LiDAR is widely used in both architectural and archeological heritage documentation, with the primary distinction between the two lying in the scale and data volume of the captured heritage. In both cases, the ultimate goal is to acquire high-precision point clouds and to rapidly generate 3D models through photogrammetric processing [103,104]. However, for occluded or hard-to-reach areas of buildings—affected by viewing angles or surrounding environments—and for fine structures such as brick carvings or intricate door panels, additional data collection using handheld laser scanners or portable photographic devices is often required [105,106].

Moreover, many historical buildings contain multiple forms of heritage—such as murals, sculptures, and complex historical structures—making cross-application of measurement techniques common. For example, MSI systems, typically employed for paintings, are increasingly used across various CH categories [107,108]. Integrating photogrammetry and MSI data enables the extraction of information far beyond that of ordinary 3D models, such as conservation status, prior interventions, natural degradation, and the influence of deposits or environmental dust [109]. Overall, the integration of multiple technologies has become the dominant trend in the construction of CH digital archives. Nevertheless, several technical bottlenecks persist in mainstream approaches. In 3D modeling, small artifacts generally involve limited data volumes, whereas massive point cloud datasets from large buildings or extensive archeological sites have become a major obstacle limiting both efficiency and scalability [110]. Since most architectural heritage is exposed to outdoor environments or remains in active use, on-site conditions—such as climate, lighting, and dust—directly affect the precision of point cloud generation [111]. In many cases, multiple LiDAR or other scanning devices must be employed to capture different building sections, necessitating data interoperability and accurate registration across varying acquisition environments [112]. Some studies have proposed voxel-based downsampling approaches, which preprocess point cloud data by segmenting the overall structure into uniform modular blocks and constraining the number of points per module, thereby reducing total data size and improving computational performance [45,113]. Others rely on 2D imagery or model reconstruction from photogrammetry to build digital archives—an approach that, while technically accessible and easy to scale, produces models not directly derived from real-world capture, thus significantly compromising their geometric accuracy and omitting critical information such as surface textures and structural layers of CH [114]. For ICH—a rapidly growing area of interest—digital preservation techniques such as motion capture and multi-camera scanning have reached relative maturity. The use of retroreflective markers attached to the performer’s body remains a mainstream approach [115]. However, most motion capture systems still introduce a certain degree of physical interference for performers, reducing the fidelity of expressive forms such as dance. Additionally, the precision of fine-grained motion capture remains limited, pointing to a key research direction for improving future CH digital archiving [116].

(2)

The implementation pathways of CH virtual exhibitions.

Providing an immersive virtual experience is the ultimate goal of CH digital archiving and serves as a crucial channel for communicating CH value and meaning to audiences. Based on differences in display media and interaction mechanisms,

Table 11. The technical path of immersive experience.

Exhibition Category	Core Technology	Display/Interaction Specifications	Typical Hardware/Software Platforms	Technical Advantage	Applicable Scenarios
Immersive space	Multi-channel projection + edge-blending system (synchronized via GPU cluster)	4–8 projectors, 2–4 K resolution per channel; FOV ≥ 270°	Barco UDX series/Christie GS series; Watchout/Unity sync server	High audience capacity Simple equipment requirements Ability to reproduce heritage scenes at a 1:1 scale	Immersive exhibition halls in museums or educational institutions, and large-scale heritage sites.
Head-mounted display	Stereo rendering engine + 6-DoF tracking (Oculus SDK/SteamVR)	90–120 Hz refresh rate; <20 ms latency	Meta Quest 3, HTC Vive Pro 2, Apple Vision Pro; Unreal Engine 5/Unity XR	High portability Rich and diverse interaction modes	Lost or inaccessible CH sites, or on-site environments enhanced with virtual information overlays.
Holographic projection	Pepper’s Ghost optical reflection/LED hologram array/LIDAR-based volumetric rendering	Image depth 0.5–1 m; brightness > 1500 nits	Holo-Stage 360; Looking Glass Portrait/Vision 3D Cube	No additional equipment required for visitors Seamless integration of virtual and physical spaces	Immersive theaters within CH institutions, light shows or architectural projections in heritage parks, and digitally augmented displays within exhibition cases.
Multimodal interaction	Integration of infrared motion capture, voice recognition, and haptic feedback modules	Latency < 30 ms; multi-user tracking ≤ 10 people	Kinect Azure/Leap Motion/Ultraleap; custom Arduino sensor network	Multi-layered and immersive experience Free from geographical and spatial constraints	Immersive galleries and thematic exhibition zones in museums or other CH institutions.

categorizes the current types of immersive exhibitions and outlines the technical principles, key features, and advantages associated with each category.

Overall, the shared objective of diverse immersive exhibition technologies is to enhance audience immersion and engagement, providing a stronger sensory and cognitive impact than that experienced at physical CH sites. At present, the application of multi-channel projection systems, various VR platforms, and on-site interaction devices has reached a relatively mature stage. For instance, immersive projection halls can be equipped with infrared or NFC sensors to track audience positions and body movements, providing corresponding visual and auditory feedback [117,118]. A key advantage of such setups is their ability to accommodate multiple users simultaneously, fostering shared virtual experiences that significantly improve visitor engagement while strengthening the educational and interpretive value of CH [119]. Notably, with the advancement of multimodal interaction technologies, ICH has emerged as one of the main beneficiaries [120]. Traditional crafts—such as pottery or weaving—have long been displayed through static exhibits or two-dimensional videos, offering limited engagement. Realistic experiences of these processes typically require physical workshops and professional instructors. Multimodal interaction technologies help overcome this barrier: for example, a mechanical arm encased in a silicone shell can replicate the tactile sensation of spinning a potter’s wheel, enabling visitors to experience the haptic and visual feedback of real-world craftsmanship [121,122]. Furthermore, by integrating VR systems or setting up multiple interactive nodes (e.g., screens, controllers, touch devices) within an exhibition space, CH virtual exhibitions can adopt gamified structures. Compared with purely sensory displays, gamification adds elements of enjoyment, co-creation, and behavioral engagement, revealing a clear research trend toward systematic experience design in future immersive CH exhibitions [123].

However, it must be emphasized that while the incorporation of multimedia content in traditional museums has become commonplace, VR and multimodal systems still face numerous challenges. First, despite over a decade of consumer-level VR availability, the development of compatible content remains a critical bottleneck. On one hand, content creation is expensive and data-intensive: reconstructing a single historical site may involve hundreds of gigabytes of raw measurements, and converting this into interactive VR content for deployment can require budgets of several hundred thousand USD, severely constraining the size of the creator community [124,125]. On the other hand, poor interoperability between different VR platforms further limits dissemination—most content is optimized for a specific device or institution, hindering cross-platform accessibility [126,127]. This mismatch between content ecosystems and hardware capabilities remains a major barrier in the VR domain. Additionally, while multimodal interaction offers the most complete and authentic CH experiences, it also involves complex hardware setups and interactive programming, making system stability and high maintenance costs persistent challenges [128]. Multi-sensor coordination requires low latency and consistent data synchronization, yet in multi-user settings, interference and tracking errors can easily occur, significantly degrading overall exhibition quality [129]. Moreover, multimodal systems are often custom-built for specific themes, and reconfiguration for new exhibitions demands considerable technical and financial resources. In summary, although immersive exhibition technologies have achieved a high level of maturity, future progress will depend on addressing key issues such as content quality enhancement, system stability improvement, and the establishment of deeper emotional connections between audiences and cultural heritage.

4.3.2. Middle Tier: System Workflow

In the previous section, we discussed the technical implementation pathways of immersive exhibitions. At the middle layer, the focus shifts toward the integration of the entire workflow—from curation to exhibition completion. Currently, the workflow of CH virtual exhibitions typically involves several interconnected stages: theme conception, content production, interaction design, platform deployment, and user access [82,130]. The curatorial concept is proposed by CH institutions or museums, which identify the exhibition theme, organize collections, and establish the narrative framework to clarify the exhibition’s objectives and structure [131]. The next step is digital content production, which—beyond the creation of digital archives discussed earlier—requires close collaboration with historical documentation sources and academic institutions [132]. For example, architectural or archeological heritage often undergoes multiple phases of discovery, preservation, and restoration, during which its cultural and historical value is gradually defined. Essential semantic information—such as architectural component attributes or the craftsmanship behind excavated artifacts—must be co-developed by archeologists, historians, and other domain experts to construct a coherent digital narrative script [133,134]. Interaction design can be divided into two layers: online exhibitions, which involve user interface layout, interaction flow, and navigation design; and offline exhibitions, which include spatial adaptation, hardware configuration, and environmental calibration [135]. Depending on curatorial objectives or institutional requirements, the exhibition can be deployed through web-based platforms, mobile applications, or digitally enhanced physical installations, forming an integrated interactive system [136,137]. To illustrate this process more intuitively, the complete workflow of CH virtual exhibitions is presented in Figure 11.

From the perspective of system-level workflow, the dual challenges of replicability and sustainability remain the primary barriers to expanding the application scope of CH virtual exhibitions. Regarding replicability, although the significant diversity among CH types makes a single universal standard impractical, the lack of data interoperability hinders the dissemination of high-quality exhibition designs and operational models [133]. For instance, different regions and museums employ a wide variety of 3D object formats—such as OBJ, FBX, glTF, and LIDO. While partial conversion between these formats can be achieved through automated tools or manual processing, essential attributes such as texture mapping, material properties, and semantic metadata often fail to transfer accurately [138,139]. In terms of platformization of digital content, major commercial VR manufacturers—including Meta, HTC, Google, and Apple—commonly use proprietary data interfaces and tie exclusive content to their devices as a competitive sales strategy. This practice effectively binds CH virtual exhibitions to specific commercial ecosystems, substantially increasing the financial burden on both curators and audiences [140,141]. As for sustainability, apart from well-established, large-scale platforms such as Google Arts & Culture, the vast majority of CH virtual exhibitions are one-off projects. In the case of permanent or “flagship” artifacts, long-term success requires continuous financial investment, technical maintenance, and content innovation by professional teams [142,143]. However, in most institutions, digital integration remains limited to basic multimedia supplements such as screens, audio guides, or static 3D models—still far from constituting a truly immersive or interactive virtual exhibition. This gap underscores the ongoing technical and financial constraints that impede the sustainable realization of CH virtual exhibition practices.

4.3.3. Top Tier: User Feedback Mechanism

An efficient and accurate evaluation of audience experience is essential for promoting the sustainable development of CH virtual exhibitions. At present, mainstream approaches to audience experience assessment generally combine subjective and objective methods. On the subjective side, questionnaires and interviews remain the most popular due to their practicality and efficiency, with analytical models such as SEM, fsQCA, and the Likert scale being commonly applied [144,145]. On the objective side, behavioral and physiological indicators are used to capture the implicit dimensions of user experience. For example, eye-tracking systems can record gaze points and fixation durations in virtual galleries, revealing users’ attentional hotspots and aesthetic preferences; meanwhile, physiological signal monitoring (e.g., heart rate, electrodermal activity, EEG) helps quantify emotional arousal and cognitive load levels [146]. In addition, interaction logs—including click patterns, navigation paths, and dwell times—offer a broader, lower-cost means of measurement, particularly suitable for large-scale or high-traffic museums [147].

It is important to emphasize that all user evaluation methods ultimately aim to build a more refined experience framework that enhances audience trust. First, given the inherently “virtual” nature of such exhibitions, achieving a high degree of perceptual authenticity—through accurate reconstruction of CH appearance, historical context, and fine details—is crucial for fostering trust and cultural attachment [148,149]. Second, the institutional credibility of the hosting platform or CH organization serves as another vital source of trust. Research has shown that factors such as the institution’s authority, the transparency of information sources, and the smoothness of interactive technologies significantly affect users’ trust formation [150,151]. Third, the ability to provide personalized content has become a dominant trend in enhancing user satisfaction. Examples include mobile apps or recommendation systems that generate individualized visit reports after tours, or offer customized visiting routes based on short pre-visit questionnaires [152,153].

Nonetheless, current audience evaluation methods still face several challenges. For questionnaire- and interview-based approaches, subjectivity in responses, cross-cultural variability, and differences in cognitive understanding among participants often undermine the reliability and interpretability of survey results [154]. For physiological measurements, although highly accurate, they require extensive data processing and specialized analytical expertise, which many CH institutions and researchers lack the technical or financial capacity to sustain. In addition, exhibition design must consider cognitive load limitations—overly complex interfaces or stimuli can fatigue visitors and diminish their motivation to explore, which is especially problematic for those engaging in leisure-oriented visits [155]. To address these issues, future improvements should focus on enhancing the cultural adaptability of exhibition design—for instance, by localizing interface language and visual style—and adopting multimodal hybrid evaluation strategies, such as combining questionnaires with eye-tracking data [40]. Furthermore, to mitigate cognitive overload, the integration of gamified design elements has emerged as a prominent trend and represents an important direction for ongoing research, which will be further discussed in the next section.

4.4. The Trend of Gamification Design

In the previous sections, the co-citation clustering analysis revealed several conclusions that align closely with the keyword clustering results; however, two emerging areas—gamified design and educational functionality in virtual exhibitions—stand out as key research trends that were not explicitly represented within the core thematic structure. Based on the reviewed literature, current research on gamification in CH virtual exhibitions has evolved into two major directions: the entertainment–education model and the education-embedded model [156,157]. The entertainment–education model refers to the educational repurposing of commercial games, in which cultural or historical learning modules are embedded within preexisting game worlds and interaction logics. A notable example is the Discovery Tour mode in Assassin’s Creed: Origins, developed in collaboration with archeologists and historians, which integrates content related to art, architecture, philosophy, and religion based on real aspects of ancient Egyptian civilization [158,159]. Although most of its elements are not derived from real-world measurement data—and the 3D models cannot be directly applied to site conservation—the game’s extraordinary level of detail and widespread critical acclaim demonstrate its immense potential for CH education and public engagement. By contrast, the education-embedded model positions cultural communication and learning as its central goals, using game mechanics primarily as tools to stimulate participation and reinforce memory. Typically initiated by CH institutions or museums, this model employs mechanisms such as mission systems, interactive puzzles, and simulation-based tasks [160,161]. For instance, in exhibitions featuring excavated artifacts, visitors may engage in simulated artifact restoration or historical scene recognition to better understand past events or craftsmanship processes. This approach maintains a serious cultural context while enhancing both participatory depth and learning engagement, making it the most prevalent form of gamified design in CH virtual exhibitions today [162].

However, a persistent tension between educational value and entertainment appeal exists in both models. Excessive emphasis on entertainment risks oversimplifying or fragmenting cultural content, thereby diminishing the exhibition’s educational integrity; conversely, overemphasis on academic seriousness may reduce engagement and increase visitors’ cognitive load, negatively affecting the overall experience [163,164]. Building upon the earlier discussion on user experience, future gamified design in CH virtual exhibitions should prioritize cultural adaptability and sustainable iteration. Exhibitions must be capable of flexible adjustment according to regional, linguistic, and audience-specific contexts, while maintaining long-term vitality through periodic updates of mission structures and content databases to ensure ongoing educational relevance [165].

Furthermore, developing a standardized design and evaluation framework represents another crucial improvement direction. Such a framework should include multidimensional indicators—covering cognitive gain, emotional engagement, and cultural understanding—to facilitate consistent and comparable assessments of user perception and learning outcomes across different gamified exhibition formats [166].

5. Discussion and Research Implications

After conducting a comprehensive analysis of the literature data, the development trajectory, research structure, and frontier themes of CH virtual exhibitions have become well defined. Moreover, the discussion of current research bottlenecks has further illuminated the directions of both technological and conceptual advancement. Building upon these insights, the following section presents predictions regarding the future development pathways of CH virtual exhibitions.

5.1. Future Research Trends

5.1.1. The Improvement and Integration of the Content Ecosystem

Based on the previous discussion of the content ecology of CH virtual exhibitions, the persistence of data silos and the lack of platform sustainability have emerged as two major bottlenecks in the field. Future research should therefore focus on developing universal content interoperability systems and fostering open, collaborative ecosystems.

First, in terms of data standards, it is essential to adopt internationally recognized formats that ensure consistent semantic structure and data openness. The CIDOC Conceptual Reference Model (CRM) serves as an effective standard for integrating CH information and documentation. It allows for advanced semantic-level data retrieval within the complex structures of CH knowledge, while also providing detailed mechanisms for defining and recording specific data points and research questions. The model was formally recognized as an ISO standard as early as 2006 [167,168].

Second, regarding platformization, Europe provides valuable reference cases. The Europeana Digital Cultural Platform of the European Union has successfully aggregated over 55 million items from more than 4000 cultural institutions across Europe through a unified metadata framework, achieving large-scale data sharing and cross-platform interoperability [169,170].

However, while broad data sharing can maximize the value of digital resources, developer ecosystem incentives and optimization mechanisms are equally vital for the sustainability of consumer-grade immersive technologies. At present, most VR content is developed for single-purpose projects or closed platforms, severely constraining long-term content viability. The recent stagnation and disappointing sales of Apple Vision Pro [171] highlight that even with strong technological and financial capabilities, the lack of openness and sustainability in the creative ecosystem remains a critical barrier to the widespread adoption of immersive content.

Therefore, future improvements should emphasize the creation of sustained developer support mechanisms, the development of lower-cost VR devices, and the reduction in platform entry barriers—for instance, enabling easy cross-platform content conversion or the integration of high-quality resources from broader databases [172,173]. While these measures may slightly compromise content exclusivity or technical refinement, expanding the user base, diversifying content offerings, and building stable and trustworthy platform ecosystems will be far more urgent priorities for the future of CH virtual exhibitions.

5.1.2. The CH in Resource-Constrained Areas Deserves More Attention

The previous national collaboration analysis revealed a pronounced regional imbalance in the digitization of CH resources. Although Europe and other developed regions exhibit a high degree of CH digital maturity and currently lead most research initiatives, this does not imply the absence of valuable CH resources in other parts of the world [174]. According to UNESCO, only about 53.6% of the global population currently has access to digital technologies, while in the least developed countries (LDCs), this rate drops to 19%. In the context of museums and virtual exhibitions, only 5% of museums in Africa and Small Island Developing States (SIDS) maintain an online presence, and the proportion capable of hosting immersive virtual exhibitions is even lower—underscoring the vast gap between these regions and developed economies such as those in Europe [175]. This disparity means that only a small fraction of the world’s population can access cross-regional CH information, while the authority to define cultural and historical values is largely concentrated within a handful of developed countries. Such imbalance weakens the diversity and ecological balance of the global CH digital domain [176]. To address this challenge, low-cost technological solutions, broader knowledge-sharing frameworks, and volunteer-based crowdsourcing models offer promising pathways [177]. For instance, after providing basic technical training to local residents, community-based digitization projects can use smartphones for photogrammetry or simple photography to document local CH assets—an approach that represents an effective model of community participation and distributed digitization [178,179].

Moreover, international assistance and cooperation are indispensable, especially in regions frequently affected by conflict, natural disasters, or humanitarian crises. In countries such as Syria, Yemen, and Iraq, professional teams have undertaken emergency documentation of war-damaged architectural heritage, employing technologies such as remote sensing and blockchain-based tracking to combat illicit CH trafficking. These initiatives help ensure that, even in the face of irreversible destruction, CH can continue to serve as a carrier of regional and national memory [175].

Overall, compared with high-cost immersive exhibition technologies, the rapid establishment of fundamental CH digital archives is a more urgent priority in resource-limited regions. This highlights the critical importance of empowering local communities, promoting open technology sharing, and fostering inclusive participation in the global digitization of cultural heritage.

5.1.3. The Transformation of Cultural Value Interpretation Brought About by Technology

Emerging exhibition technologies such as VR, AR, haptic interfaces, and multimodal interaction systems are profoundly transforming how audiences perceive and interpret the value of CH. These technologies provide immersive and interactive experiences that reconfigure traditional exhibition narratives. For example, AR technology enables the superimposition of digital images, audio, and 3D reconstructions onto real-world artifacts or heritage sites, allowing visitors to engage with multi-layered narrative content in situ. Invisible historical layers—such as the evolution of an architectural heritage site—can thus be dynamically visualized, offering audiences a richer understanding of cultural meaning [180]. Multimodal systems also help create emotionally resonant, immersive environments that evoke empathy and embodied understanding. This sense of experiential authenticity has become a key criterion for evaluating exhibition effectiveness: visitors now value not only historical accuracy but also the emotional connection and situational resonance that contribute to a deeper sense of realism [181,182].

However, while these technologies empower new forms of cultural interpretation, they also raise critical questions of authenticity and ethics [183,184]. On one hand, hyper-realistic virtual reconstructions may blur the line between reality and imagination, leading audiences to mistake digitally recreated scenes for historical fact. On the other hand, issues of interpretive authority emerge: Who determines which version of history a virtual exhibition presents? Which narratives are emphasized or omitted? Such questions reveal the risk of cultural bias in digital storytelling [185,186]. Several emerging solutions in the field offer meaningful approaches to these dilemmas. One strategy is the adoption of layered narrative design, in which interfaces clearly differentiate among original data, archeological hypotheses, and creative reconstructions—ensuring visitors understand what constitutes verified history versus interpretive addition. This approach allows audiences to enjoy rich content without compromising historical authenticity. Another strategy involves multi-user and multi-perspective presentation, where virtual exhibitions integrate diverse narrative voices—including those of local communities, scholars, and visitors—to co-construct and recontextualize cultural meaning [182,187]. Such polyphonic curatorial models promote a balance between technological empowerment and ethical responsibility, fostering inclusivity and acknowledging the participation and interpretive rights of different stakeholder groups in shaping the digital futures of cultural heritage.

5.1.4. The Application Potential of AI and Machine Learning (ML)

AI and ML hold tremendous potential for application in CH virtual exhibitions, primarily across three domains: curation, visitor guidance, and content generation.

First, in intelligent curation, AI can employ knowledge graphs and topic clustering algorithms to identify relationships within massive collections and automatically generate exhibition concepts. For example, AI can analyze museum collection data, identify latent thematic connections among artifacts, draft interpretive texts and wall labels, and then allow curators to review and refine the final content [188]. Compared with traditional manual organization, AI can process vast amounts of historical data and link cross-temporal themes, making it particularly advantageous for large-scale collections and capable of revealing previously overlooked historical relationships [189]. This not only drives innovation in exhibition design but also creates new opportunities for reinterpreting historical narratives and regional cultures. Nevertheless, it is important to emphasize that AI-assisted curation still requires significant human oversight, as issues of academic accuracy, narrative ethics, and explainability remain major concerns [190].

Second, AI can enhance personalized visitor guidance, creating tailored experiences for individual audiences. By modeling user behavior—such as the exhibits viewed, dwell time, and preferred themes—recommendation algorithms can dynamically adjust exhibition content and provide context-specific explanations and visiting routes [191,192]. For instance, the ArtLens application developed by the Cleveland Museum of Art integrates a digital gallery, venue navigation, and interactive exhibits. Serving as the museum’s central digital interface, ArtLens also mitigates information overload in major exhibition spaces by generating personalized tour recommendations [193]. Looking forward, delivering customized tours and post-visit analytical reports is expected to become a defining innovation for future CH institutions and museums.

Finally, generative AI demonstrates transformative potential in heritage restoration. Deep learning models can reconstruct lost or damaged heritage assets using limited historical resources [194,195]. A collaborative study between Japan and China, for example, showed that by training a neural network on pre-damage photographs of a bas-relief, the model was able to reproduce up to 95% of the original details and depth accuracy, greatly outperforming traditional manual reconstruction methods [196]. Furthermore, integrating AI, ML, UAVs, LiDAR, and historical datasets enables the virtual reconstruction of large-scale heritage sites—such as ancient Pompeii—whose complexity and scale make manual modeling nearly impossible. This convergence of technologies underscores the vast potential of AI and ML in the preservation, reconstruction, and reinterpretation of cultural heritage in the years to come.

5.2. Research Limitations

This study has several limitations in terms of data collection and analytical procedures. First, the exclusive use of the Web of Science Core Collection (WoS CC) as the sole data source—although validated by numerous prior studies—may have resulted in the omission of innovative or distinctive works indexed in broader databases such as Scopus. Second, only English language publications were included, which may have excluded relevant studies and journals published in non-English regions or languages. Finally, despite employing three analytical tools, the analyses were necessarily constrained by the preset functionalities and frameworks of these software packages. Future research should explore more diverse and efficient analytical methods or adopt alternative perspectives to uncover deeper insights and identify emerging trends and iterative research paradigms within the field.

6. Conclusions

This study conducted a systematic review and bibliometric analysis of literature on CH virtual exhibitions published between 1999 and 2025, providing comprehensive responses and integrated interpretations of the six key research questions.

(1) Developmental stages of the field:

The evolution of CH virtual exhibitions can be divided into three stages: the concept formation stage (1999–2013), the technological expansion stage (2014–2018), and the integration and innovation stage (2019–2025). Early research focused on conceptual construction and experimental demonstrations; the middle phase experienced rapid growth with the popularization of VR/AR, 3D modeling, and multimedia visualization; and the recent phase has seen integrated development across multimodal interaction, educational incorporation, and narrative innovation. Notably, the global digital transformation triggered by the COVID-19 pandemic has accelerated the institutionalization of virtual exhibitions, transforming them from supplementary displays into essential modes of cultural communication and public education.

(2) National policies and regional disparities:

Government policies and national digital strategies have had a substantial impact on research activity and influence within this field. China and Italy lead in publication volume, supported by state-led digital heritage initiatives, while Australia stands out in citation impact due to its long-standing commitment to cultural digitization and research infrastructure. These findings suggest that policy continuity and institutional investment remain key drivers for the maturation and internationalization of CH virtual exhibition research.

(3) Research networks and disciplinary structure:

Author and institutional collaboration networks reveal strong interdisciplinary integration, connecting heritage science, information science, and computer visualization into a coherent knowledge framework. This structure reflects the research transition from technological implementation to cultural exchange and cognitive interpretation.

(4) Keyword structure and knowledge organization:

Keyword clustering reveals a three-tier knowledge framework in CH virtual exhibitions: the technology-driven tier (e.g., VR/AR, LiDAR, photogrammetry), the application-scenario tier (e.g., exhibition workflows, heritage tourism, and educational communication), and the user-experience tier (e.g., audience engagement, trust mechanisms, and gamified participation). Together, these tiers form a dynamic “technology–application–experience” feedback loop, illustrating the field’s progression from technological innovation to value-oriented cultural communication.

(5) Research hotspots and core themes:

Co-citation and thematic clustering analyses identify digital strategies and museum transformation (Cluster 0), user behavior and acceptance mechanisms (Cluster 2), and immersive system design integrating educational functions (Cluster 7) as the most active research directions. Technological innovation, user engagement, and educational dissemination exhibit synergistic evolution, forming the core momentum and convergence point of CH virtual exhibition research.

(6) Emerging changes in cultural interpretation:

New exhibition paradigms—such as AI-assisted curation, education-oriented gamification, and multisensory storytelling—are redefining public interaction with heritage. Cultural interpretation is shifting from static display to participatory experience, emphasizing the integration of emotion, education, and social meaning. However, these developments also raise issues of authenticity and ethics, which require transparent curatorial practices and culturally adaptive interpretive mechanisms.

Overall, CH virtual exhibitions have transitioned from a technological experimental phase to a mature interdisciplinary system. To achieve a more inclusive and sustainable trajectory, future research should deepen its focus in three areas:

First, building an open and interoperable content ecosystem. Current bottlenecks lie in data silos and inconsistent standards; future efforts should promote cross-institutional content sharing and long-term preservation through semantic interoperability standards such as CIDOC CRM and the FAIR principles. Collaborative data platforms can enhance transparency and public participation, transforming digital heritage resources into reusable social assets.

Second, empowering community-driven digitization and technology sharing in resource-limited regions. Developing countries often possess rich yet fragile heritage assets; digitization should prioritize local participation through community training, low-cost modeling (e.g., mobile photogrammetry), and regional partnerships. When communities themselves become the recorders and interpreters of heritage, digital outcomes can more authentically reflect their cultural contexts. International organizations and research institutions should further advance knowledge sharing, equipment support, and capacity-building initiatives to achieve cultural digitization that is both equitable and locally sustainable.

Third, establishing ethical and interpretive frameworks suited to cultural contexts amid rapid AI and immersive technology development. Although AI-driven curation and content generation can enhance efficiency and personalization, they also risk narrative bias and authenticity ambiguity. Future studies should prioritize algorithmic transparency, cultural adaptability, and multi-voiced storytelling mechanisms to ensure that technological innovation continues to serve cultural understanding and social responsibility.

By systematically addressing the six research questions and proposing forward-looking directions, this study constructs an integrated analytical framework linking technological evolution, institutional dynamics, community practice, and cultural interpretation, providing both theoretical support and a global perspective for the academic advancement and practical transformation of CH virtual exhibitions.