There Is a Digital Art History

Summary Over the past decade humanities researchers have increasingly come to embrace digital methods. Art historians, however, have often resisted engaging with these developments. In this article, we explore the driving factors behind art history's reticence toward the digital turn in the humanities. Reflecting on the historiographic trajectory of the emerging field of digital art history (DAH) versus art history more generally, we selected a sample of recent articles published between 2010-2019. We used a mixture of methods, both digital and non-digital, to uncover the prevalence for different art-historical theories in DAH versus mainstream art history. We began our study by performing a text mining analysis on the references and bibliography of articles published in DAH, Art Journal, and Art History. Once we had determined a list of frequently-cited authors, we dug deeper to see how they were discussed in the body of individual texts. In other words, we employed traditional humanities methods: close reading and interpretation. DAH is typically positioned as something completely new to the discipline. However, as this study shows, DAH is closely tied to particular pre-digital methods and theories of art history, namely formalist and iconographic methods prevalent during the late nineteenth and early twentieth-century, rather than critical theory methods commonly found in more recent art-historical scholarship. Based on our analysis, we argue that DAH methods have not been embraced by art historians more generally because of fundamental differences in the theoretical underpinnings of DAH versus the broader field of art history.

This article responds to two issues affecting the field of contemporary art history: digital technology and the so-called computational turn in the humanities. It is divided into two parts: the first connects problems with “digital art history,” an offspring of digital humanities, to neoliberal metrics; the second suggests how digital art history’s“distant reading” might nevertheless be deployed critically in the analysis of contemporary art.

Integration of information and communication technology platforms is one of the incentives for the development of Digital Humanities (DH) projects abroad in their increasingly close interaction with academic, primarily University libraries. The purpose of the article is to analyze the deepening interaction of libraries with DH projects within the single digital paradigm, as well as to analyze new services for project management, data curation, organization of online publications, preservation and exchange of databases, etc. In the era of digital convergence, large scientific libraries abroad are actively developing new models of cooperation with DH projects, forming special service packages for continuous servicing of scientists on the basis of the integrated approach.The article discusses the new concept of library services for DH projects, draws attention to the problem of classification and ranking of library services in order to streamline the work and improve its efficiency, describes the existing multi-level service models developed in the library systems of the major US Universities. In the context of formation of the new package of library services within the framework of DH projects, the article considers the new practices and operational models on which they are based. The author describes the experience of University libraries in the Netherlands and Canada. Under the innovation pressure in the field of library technologies, the service package offered by scientific libraries in developed countries is constantly expanding and modernizing. It includes not only digitization of printed materials, computer assistance, but also management of databases, preservation and exchange of scientific data, organization of digital publications, consultations in the field of copyright protection, etc. Development of new practices in the field of digital services management changes the organizational paradigm of library work, puts forward new standards of professional activity, which require continuous training to meet, i.e. launches an institutional rebranding of the library.

Over the last two decades, teaching, learning and research in higher education have developed a growing digital presence. Digital development in the humanities has been slow relative to most other areas in academia, and with some exceptions art and art history have enjoyed slow digital growth within the humanities. Within this environment, the article here presents one collaborative model for digital art history, rare in its exclusive focus on undergraduate “junior scholars”. Undergraduate senior-level art history and studio art students at Providence College collaborate annually with art history and studio art faculty to publish their senior theses in print format as the Art Journal. In the last few years, students, faculty and digital library staff have enhanced this collaboration to include the publishing from process to product of the Art Journal as a complementary digital Art Journal. They collaborate in creating digital art history and digital studio art in order to bring exponentially greater meaning, significance and visibility to the students’ senior culminating works through real-world digital publishing, including quality control, copyright issues and ideas related to persistent access and ongoing global visibility for the scholarly and creative works and for the student scholars. These students function as real-world collaborative scholarly partners in publishing their culminating academic and artistic work globally and persistently accessible in Providence College’s digital repositories. This case study evidences engagement in meaningful digital knowledge creation focused on the intellectual and creative output of student-scholars and student-artists (art historians and studio artists) as a model for other student-faculty-digital library professional collaborations.

Reviewed by: Empires of Vision: A Reader ed. by Martin Jay and Sumathi Ramaswamy Carla Manfredi Empires of Vision: A Reader. Edited by martin jay and sumathi ramaswamy. Durham, N.C.: Duke University Press, 2014. 688 pp. $119.95 (cloth); $32.95 (paper). Empires of Vision contributes to our understanding of the visual cultures of European imperialism and to their afterlives in postcolonial milieus. Thoughtfully edited by cultural historians Martin Jay and Sumathi Ramaswamy, the volume includes twenty-one reprinted essays and is divided into two sections: “The Imperial Optic” and “Postcolonial Looking.” The first section introduces five visual media within a trans-historical and global context that spans roughly five hundred years of imperialism. The second section focuses on the decades following decolonization and explores the different ways in which subaltern artists have responded to or “look[ed] back” (p. 3) at Europe. The selections offer a multitude of historical, regional, cultural, and theoretical perspectives, which will be of interest to an interdisciplinary readership. Although many of the volume’s selections will undoubtedly be familiar to seasoned scholars, they will provide graduate students with a solid grounding for future historical and theoretical research across the fields of imperialism, postcolonialism, visual and cultural studies, and art history. It is impossible to do justice to any book in a brief review let alone to extracts from twenty-one impressive pieces of scholarship. Thus, this review focuses on the editors’ contributions to Empires of Vision, which are bold reminders of what cross-disciplinary study can achieve: a recasting of conventional, written histories and a questioning of entrenched theoretical paradigms. In “The Work of Vision in the Age of European Empires,” Ramaswamy offers a sophisticated and lucid introduction to current critical approaches to colonialism and visual culture. The stakes of Empires of Vision are clearly defined: “The image is a site where new accounts of empire, the (post)colony, and Europe itself emerge and depart from—even challenge—the more familiar narrative line(s) of nonvisual histories” (p. 3). The volume, explains Ramaswamy, responds to and encourages a gradual academic reorientation. Despite the appearance of several recent visual culture anthologies, the study of art remains largely limited to art history; indeed, histories of colonialism and postcolonialism marginalize the integral role of visual culture in the production of colonial violence. Ramaswamy supports her claim with an anecdote about Edward Said, who confessed that “‘just to think about the visual arts generally sends me into a panic’” (p. 5). Panic becomes a productive metaphor for theorizing how images are often “disorderly” and “unpredictable” and sometimes “incoherent”; thus, Empires of [End Page 927] Vision offers alternative histories to those presented in official textual archives (pp. 5–6). If colonial and postcolonial studies have given short shrift to the visual arts, then visual culture studies also need to catch up. In conventional visual culture accounts, the presence of empire is still obscured, resulting in a European framework of visuality (p. 10). To illustrate this argument, Ramaswamy refers to a compelling 1996 “Visual Culture Questionnaire” published in the journal October: No one who responded to this questionnaire was a specialist in visual arts from regions other than Europe or the United States (pp. 10–11). Needless to say, the questionnaire highlights the importance of this collection, which will militate against parochialism (whether disciplinary or otherwise) and encourage interdisciplinary and interregional approaches to visuality and empire. In their introduction to part 1, “The Imperial Optic,” Ramaswamy and Jay trouble Eurocentric approaches to visual culture by underscoring the dynamic interactions between metropolitan technologies and indigenous traditions and practices. Thus, the “mutually constitutive relationship between empire and image-work” engendered new practices of seeing in and outside Europe (p. 25). Part 1 is divided into four sections, which deal with easel painting, mass-printed illustrations, cartography, and photography/film respectively. Anthropologist Deborah Poole’s influential notion of “visual economy” (1997) serves as an underlying concept for this part. The term “visual economy,” rather than “visual culture” is better suited to consider how images circulate within wide social, cultural, geographical, and imperial networks. In other words, individuals do not have to share a common (visual) culture in order to belong to the...

This article poses a critical approach to the emerging field of study that is being denominated “digital art history.” I discuss the kinds of digital processes experienced by art history as a discipline, the conflicts and challenges implicit in this new working environment, and I propose a systematic, programmatic critical discourse that will examine the contradictions, paradoxes, inconsistencies, and marginalization underlying the digital evolution of art history.

The Summer School on Digital Art History (DAHSS) is an ongoing joint initiative of the University of Malaga and the University of California, Berkeley.

This paper reflects on the recent collaboration between the National Gallery Research Centre and the Department of Digital Humanities at King's College London (NG/DDH). Using the stock books located in the archives of the art dealers Thos. Agnew & Sons as an example, NG/DDH sought to address the potential that such archival sources might have in the world of digital research and Digital Asset Management. Building on this research, NG/DDH participated in the Getty Advanced Workshop on Network Analysis and Digital Art History, which encouraged us to think particularly about the role and nature of network analysis in the field of Digital Art History. In this paper we briefly describe the technical process of digitization, and the possibilities opened up by converting the transactional information contained in the stock books into structured data. We then set this in the historical context of network structures, as embodied by the World Wide Web and social media, and especially the distinction that these types of network have entrenched between objects (or persons) and topics (or other abstract entities). We conclude with an argument that problematizing this distinction – which is enshrined in the whole concept of networks and network analysis – is key if this method is to be applied to art history research.

What happened after the Art history in digital dimensions report of 2017? To what extent have the recommendations been implemented? Has digital art history itself changed in this period? Are the challenges and recommendations outlined in the report still pertinent? What is the role of or is there a role in digital art history for the art librarian in an academic or museum setting?

The chapter provides a brief overview of the most important advancements in the application of quantitative methods in art history, followed by the description of research conducted between 2014 and 2018 at the project ARTNET.

The sustainability of digital humanities research projects is a pressing issue for humanities computing. Currently, even well-established large digital projects like the Linguistic Atlas Project (LAP) are at future risk because funding and other resources are contingent on grant funding or faculty status of the director, neither of which will necessarily be available to maintain the project over time. The mission of the university library, however, includes archiving and dissemination, now increasingly of digital materials as well as traditional paper. Collaboration with the university library is the only realistic option for long-term sustainability of digital humanities projects in the current environment. Unlike paper collections, which only require secure storage, digital projects also require the means of adaptation to new electronic media and operating environments. Even data storage requires that materials from digital projects be included in library media refresh cycles, which will include transfer of old data to new media as technology develops. Projects like LAP should provide resources to assist the library in starting the project archive, including staff time, and funding for equipment. Project metadata must be provided and, to the extent possible, integrated with library systems and finding aids. Project staff will also need to maintain a Web presence and tools developed for the project. Such cooperation leads toward the development of a digital institutional repository, in which research results and tools may be maintained in the library, not just for the humanities but across many disciplines.

Reviewed by: Digital Humanities in the Library: Challenges and Opportunities for Subject Specialists ed. by Arianne Hartsell-Gundy et al. Harriett E. Green Digital Humanities in the Library: Challenges and Opportunities for Subject Specialists, ed. Arianne Hartsell-Gundy, Laura Braunstein, and Liorah Golomb. Chicago: Association of College and Research Libraries (ACRL), 2015. 291 pages. $68.00 (ISBN 978-0-8389-8767-4) From early on, academic libraries and library and information science programs have pioneered digital humanities (DH) projects and initiatives. These have included the Valley of the Shadow Project, a digital history project depicting the experiences of Confederate soldiers in the Civil War; the Text Creation Partnership, which produces digital editions of early print books; Documenting the American South, a collection of full-text digital sources on Southern history and culture; and the Walt Whitman Archive, which has writings and other materials about the American poet. Yet only in recent years has the role of librarians noticeably shifted from facilitating access to content to taking a more actively engaged role by collaborating with faculty and initiating DH projects. And this shift is necessary. As Tyler Walters and Katherine Skinner observe in New Roles for New Times: Digital Curation for Preservation (Chicago: ACRL, 2011, p. 72), “Libraries must also be concerned about losing ground within their campus environment by not meeting the digital needs of the scholarly community. If we do not seek to engage with the digital humanities, other entities will.” Featuring firsthand case studies of academic librarians engaging in DH research and teaching, Digital Humanities in the Library: Challenges and Opportunities for Subject Specialists explores the strategies necessary for taking a more active role in DH projects. This volume is divided into four sections: “Reasons for Subject Specialists to Acquire DH Skills”; “Getting Involved in Digital Humanities”; “Collaboration, Spaces, and Instruction”; and “Projects in Focus: From Conception to Completion and Beyond.” Chapters in each section describe in detail diverse projects, collaborations, and initiatives undertaken by librarians over the past decade. Practical in perspective and scope, the book pays particular attention to teaching and learning, providing case studies of collaborations between librarians and faculty in small and large colleges and universities. Other chapters cover strategies for outreach and engagement with stakeholders, researchers, and organizations. Still others offer insightful “how-to” approaches for building infrastructure to support DH scholarship, including digital scholarship centers and learning spaces, and for outreach and sustainability of library-centered digital [End Page 207] projects. The most impactful contributions in Digital Humanities in the Library cast a critical eye on the theory and practice of digital humanities and the role of libraries therein. These include Caro Pinto’s chapter “Construction and Disruption: Building Communities of Practice, Queering Subject Liaisons,” in which Pinto exhorts librarians to “disrupt toward solidarity and innovate toward communities of practice in digital humanities.” Another stimulating chapter is Liorah Golomb’s account of her own text mining project, “Dipping a Toe into the DH Waters: A Librarian’s Experience,” which illustrates the ways in which some librarians are already digital scholars. (p. 49) Many of the lessons in Digital Humanities in the Library could also apply to digital scholarship in the social sciences and sciences. In addition to providing a good overview of current library activity in digital humanities, this volume will be beneficial to librarians looking to become involved in DH and unsure about where to begin. Addressing a wide range of institutions, skills, and projects, Digital Humanities in the Library will help librarians explore the dynamic world of the digital humanities. Harriett E. Green University of Illinois at Urbana–Champaign Green19@illinois.edu Copyright © 2015 The Johns Hopkins University Press

Click to increase image sizeClick to decrease image size Notes1. The Febuary 2011 annual meeting of the College Art Association (CAA) in New York had an entire session organized by Patricia Mainardi dedicated to “The Crisis in Art History.” The papers were later published in a special issue of Visual Resources (27, no. 4, 2011). Yet the present crisis in the discipline can be considered just the latest of many that it has faced at least since the 1970s, when a whole new range of theoretical references, such as postcolonial theories, psychoanalysis, gender studies, and semiotics, pressured art historians to reflect about their specific approaches to art. Hans Belting's Das Ende der Kunstgeschichte? came out in 1983, centering its fire on the power structures of the field in Germany, and in 1982 the Art Journal also produced an issue on “The Crisis in the Discipline,” edited by Henri Zerner, with articles by such leading art historians as Rosalind Krauss, David Summers, and Donald Preziosi, among others. At that time, the main topic of discussion was the incorporation of new objects of study in the art historical field, such as advertisement, photography, and other aspects of visual culture, bringing art history closer to so-called visual studies. Concerns about developing a global art history, something that still resonates today, appeared at the turn of the century with important publications such as Hans Belting's Bild-Anthropologie (2001) and David Summers's Real Spaces (2003). Both books demonstrate a special interest in anthropology, which persists today with the incorporation of anthropological theories such as those of Alfred Gell and Bruno Latour, for example, into the field. Belting, Das Ende der Kunstgeschichte? (Munich: Deutscher Kunstverlag, 1983), trans. as The End of the History of Art? (Chicago: University of Chicago Press, 1987); Zerner, ed., “The Crisis in the Discipline,” special issue, Art Journal 42, no. 4 (Winter 1982); Belting, Bild-Anthropologie (Munich: W. Fink, 2001); and Summers, Real Spaces: World Art History and the Rise of Western Modernism (London: Phaidon Press, 2003).2. This was the 33rd Congress of the Comité International d’Histoire de l’Art, Germanische Nationalmuseum Nuremberg, July 15–20, 2012. Its theme, “The Challenge of the Object,” appears somewhat conservative when considered from the perspective of contemporary art. Since at least the 1960s Conceptual art has done away with the idea of a necessary connection between art and the material world, which had important political consequences as well.3. The same logic that we see here organizes much of the “globalized” art world. In a review of the exhibition Africa Remix: L’art contemporain d’un continent, presented at the Centre Georges Pompidou, Paris, in 2005, Roberto Conduru notes that African artists still need to pass through the established institutions of art in the United States or in Europe to enter the circuits of the international market and art world. “With one sole exception—Wim Botha … all of the remaining 87 artists already represented their country and/or the Continent in the Biennials that proliferate throughout the world, and also presented their work in prestigious institutions within the system of international art.” Conduru, “O mundo é uma tribo,” Concinnitas (UERJ, Rio de Janeiro), no. 8 (July 2005): 198–202.4. Although this in fact is a very significant way to approach Brazilian visual culture, it should not be exclusive. Important work has been done in the last few years on the process of the transfer and circulation of material culture within the Portuguese Empire and in Brazil. See Jens M. Baumgarten, “Transformation asiatischer Artefakte in brasilianischen Kontexten,” in Topologien des Reisens, ed. Alexandra Karentzos, Alma-Elisa Kittner, and Julia Reuter (Trier: Universität Trier, 2009), vol. 1, 178–94.5. A good example can be found in Rodrigo Naves's interpretation of Jean-Baptiste Debret's Brazilian period: “Definitely, the existence of slavery hindered once and for all any effort to transplant the classical form as truth into Brazil.” Naves, A forma difícil: Ensaios sobre arte brasileira (São Paulo: Ática,1996), 71.6. Moema is the tragic character of Frei Santa Rita Durão's epic poem Caramuru. She was the sister of Caramuru's wife, Paraguaçu, and drowned while swimming after the ship that carried Caramuru back to Europe.7. Regarding the position that Baroque occupies in Brazilian art history, it is interesting that in the first Pan-American Congress of Architects, organized in Uruguay in 1920, Alexandre de Albuquerque proposed the Baroque manner of Minas Gerais as the source for a Brazilian national style, since, according to him, Brazil did not possess a strong indigenous culture, as did other countries in South America.8. Sergio Miceli, Nacional estrangeiro (São Paulo: Companhia das Letras, 2003).9. The study of art history is relatively recent in Brazil, and by examining other contributions to the history of culture in the country, one can find elements that could help to impel art history in new directions. The writer Mario de Andrade, one of the main protagonists of Brazilian modernism, for instance, developed a genuine interest in African and Native culture and in 1936 wrote a project for the creation of an Institute for National Patrimony in which he included material and immaterial culture of the different populations in Brazil. In the same period, the sociologist Gilberto Freyre, who had studied with Franz Boas, wrote his classic book Casa-grande e senzala, in which he examines the diversity of contributions to Brazilian culture. In the 1940s and 1950s, Lina Bo Bardi organized exhibitions in Bahia and at the São Paulo Art Museum (MASP) based on her research on popular art in Brazil. See, among others, Antonio Gilberto Nogueira, Por um inventário dos sentidos: Mário de Andrade e a concepçã de patrimônio e inventário (São Paulo: Hucitec/Fapesp., 2005); Marina Grinover and Silvana Rubino, eds., Lina por escrito: Textos escolhidos de Lina Bo Bardi (São Paulo: Cosac e Naify, 2009); and Gilberto Freyre, Casa-grande e senzala (São Paulo: Global, 2003).10. “From this we conclude that … art, in its simple forms, is not necessarily expressive of purposive action, but rather based upon our reactions to forms that develop through mastery of technique.” Franz Boas, Primitive Art (New York: Dover, 1955), 62. I also think that Boas's question about what could be considered an aesthetic object in these cultures is a very important one because it definitely secures these objects as of interest for the field of art history, at the same time that it forces us to listen to the point of view of their creators.11. On Assuriní visual culture, see Regina Polo Müller, Os Assuriní do Xingu: História e arte (Campinas: Unicamp, 1993). When dealing with contemporary indigenous culture we must also acknowledge the impact of the market on how aesthetic value is assigned. This, of course, implies recognizing that indigenous culture is in fact alive and interacting with other sectors of society today, resisting the tendency of both anthropologists and art historians to imagine these cultures as if “frozen in the past.”12. Tapirage is the name given to a technique used by some indigenous groups to alter the feather colors of live birds. See Amy Buono, “Crafts of Color: Tupi Tapirage in Early Colonial Brazil,” in The Materiality of Color: The Production, Circulation, and Application of Dyes and Pigments 1400–1800, ed. Andrea Feeser, Maureen Daly Goggin, and Beth Fowkes (Aldershot, U.K.: Ashgate Press, 2012), 8–40.13. Amy Buono, “Indigeneity as Corporeality: The ‘Tupinambization’ of the Early Modern Atlantic,” in Art and Its Histories in Brazil, by Claudia Mattos and Roberto Conduru, Issues and Debates (Los Angeles: Getty Research Institute, forthcoming).14. On this issue, see Joanne Rappaport and Tom Cummins, Beyond the Lettered City: Indigenous Literacies in the Americas (Durham, N.C.: Duke University Press, 2012), 1–26.15. As I have argued above, there cannot be one unitary Brazilian art; given the constant dynamic exchange between local and transferred material and immaterial culture, we must substitute this concept for a more dynamic view of the arts in Brazil.16. Terreiro is the name for buildings and spaces that contain Afro-Brazilian religious activities. Quilombos are historic settlements of fugitive and freed slaves that were created in remote places within the interior of Brazil.17. We must learn to see the circulation of people, ideas, and objects between different localities in Africa and Brazil as a two-way process. As Roberto Conduru notes, although slavery implied the forced migration of a much larger population from Africa to the Americas, there was also some significant immigration in the other direction. This is well documented, for example, in the development of agudá architecture in the region of the Gulf of Guinea (Togo, Benin, and Nigeria). Conduru, “Sobrevivência e invenção—conexões artísticas entre Brasil e África,” in Mattos and Conduru, Art and Its Histories in Brazil.18. The direct relation between art and art history has been pointed out numerous times. Wolfgang Kemp (“Benjamin and Aby Warburg,” Kritische Berichte, no. 3 [1975]: 5–6) has argued that Dadaist collage played an important role in the development of Aby Warburg's Mnemosyne Atlas, for instance. See Kurt W. Forster, “Die Hamburg-Amerika-Linie, oder: Warburgs Kunstwissenschaft zwischen den Kontinenten,” in Aby Warburg: Akten des internationalen Symposions, Hamburg, ed. Horst Bredekamp, Michael Diers, and Charlotte Schoell-Glass (Weinheim: Wiley-VCH Verlag), 11–37. Performance art has helped us understand the theatricality of Baroque art, and it also can be very productive in interpreting much of what we call Native art, for instance.19. Thierry Dufrêne, “Pour en finir avec le corset de la chronologie: Vers une histoire de l’art élargie qui rapproche les temps et les oeuvres,” in Théâtre du Monde, exh. cat. (Paris: FAGE, 2013), 30–37. The exhibition at La Maison Rouge, which showed the collection of eccentric Tasmanian collector David Walsh, was curated by Jean-Hubert Martin, who also used Surrealist-inspired anachronistic models to display the objects.20. Belting, Bild-Anthropologie; and Summers, Real Spaces. For a critique of David Summers's book in relation to the question of a new world art history, see James Elkins, “On David Summers's Real Spaces,” in Is Art History Global? (New York: Routledge, 2007), 41–72.Additional informationNotes on contributorsClaudia MattosClaudia Mattos, professor of art history at the University of Campinas, earned a PhD from the Freie Universität, Berlin, in 1996. She publishes on nineteenth- and twentieth-century art and is preparing one book on art and ecology in Brazil and editing another on Brazilian art history for Getty Publications [Instituto de Artes, Departamento de Artes Plásticas, Rua Elis Regina 50, Cidade Universitária Zeferino Vaz, 13083-854, Campinas, SP, Brazil, cvmattos@iar.unicamp.br].

Abstract— Audio data carries rich information in the form of speech, music, and environmental sounds, but its raw waveform is often complex and high-dimensional, making direct analysis difficult. Neural network architectures have emerged as powerful tools for extracting meaningful representations from audio signals, enabling efficient analysis and interpretation. Recurrent Neural Networks (RNNs), and Transformer-based models— for learning robust and discriminative features from audio. By automatically capturing temporal, spectral, and contextual patterns, these architectures significantly improve performance in tasks such as speech recognition, speaker identification, music classification, and environmental sound detection. The findings highlight the potential of neural networks to replace traditional handcrafted features, thereby advancing the development of scalable, accurate, and realtime audio processing applications. The rapid growth of audio data across domains such as speech, music, healthcare, and environmental monitoring has created a strong need for effective methods to extract meaningful representations from complex audio signals. Traditional approaches rely on handcrafted features like MFCCs and spectrogram descriptors, which often fail to capture the full temporal and spectral dynamics present in raw audio. Keywords: Neural Networks, Deep Learning, Audio Representation Learning, Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Transformer Models, Feature Extraction, Speech Recognition, Speaker Identification, Sound Event Detection, Spectrogram Analysis, Audio Signal Processing

In recent years, the transformer model has been widely used in computer-vision tasks and has achieved impressive results. Unfortunately, these transformer-based models have the common drawback of having many parameters and a large memory footprint, causing them to be difficult to deploy on mobiles as lightweight convolutional neural networks. To address these issues, a Vision Transformer (ViT) model, named the lightweight Vision Transformer (LW-ViT) model, is proposed to reduce the complexity of the transformer-based model. The model is applied to offline handwritten Chinese character recognition. The design of the LW-ViT model is inspired by MobileViT. The lightweight ViT model reduces the number of parameters and FLOPs by reducing the number of transformer blocks and the MV2 layer based on the overall framework of the MobileViT model. The number of parameters and FLOPs for the LW-ViT model was 0.48 million and 0.22 G, respectively, and it ultimately achieved a high recognition accuracy of 95.8% on the dataset. Furthermore, compared to the MobileViT model, the number of parameters was reduced by 53.8%, and the FLOPs were reduced by 18.5%. The experimental results show that the LW-ViT model has a low number of parameters, proving the correctness and feasibility of the proposed model.