SUMAC 2024

6th workshop on analysis, understanding
and promotion of heritage contents

Advances in machine learning, signal processing,
multimodal techniques and human-machine interaction

In conjunction with ACM Multimedia 2024
28 Oct - 1 November, 2024, Melbourne (Australia)


News and Updates


Overview

The sixth version of the SUMAC (analySis, Understanding and proMotion of heritAge Contents) workshop, like its predecessors, focuses on analyzing, processing and valorizing all types of data related to cultural heritage, including tangible and intangible heritage. As stated by UNESCO, cultural heritage provides societies with a wealth of resources inherited from the past, created in the present for the benefit of future generations.

Digital heritage data acquired are naturally massive and address a large diversity of monomodal modalities (text, structured referentials, image, video, 3D, music, sensor data). Their processing and promotion put into light several scientific challenges as well as various new use cases that are of topical interest today for the ACM Multimedia community, both for academics and industries. Like in the previous editions, we will strive to value the sharing of knowledge, algorithms and experiments; and also open source software and open data, by encouraging the submission of articles that promote this sharing policy.

Abundant heritage data is available in the most recent years. Older data, that can be called the big data of the past, are mostly locked -- they currently remain largely “hidden” from the public, in galleries, libraries, archives, museums or data producers' infrastructures. Processing heritage data to increase their visibility will act as a game changer and contribute to a large panel of communities, by enabling an outstanding pool of inter-operable data, not only as a service to citizens but also to public or private actors, by challenging the research methods at the crossing of computer science, artificial intelligence and digital humanities.


Call for Papers

The ambition of SUMAC is to bring together researchers and practitioners from different disciplines to share ideas and methods on current trends in the analysis, understanding and promotion of heritage contents. These challenges are reflected in the corresponding sub-fields of machine learning, signal processing, multi-modal techniques and human-machine interaction. We welcome research contributions for the following (but not limited to) topics:

  • Monomodal analysis: image, text, video, 3D, music, sensor data and structured referentials
  • Information retrieval for multimedia heritage
  • Automated archaeology and heritage data processing
  • Multi-modal deep learning and time series analysis for heritage data
  • Heritage modeling, visualization, and virtualization
  • Smart digitization and reconstruction of heritage data
  • Open heritage data and bench-marking

The scope of targeted applications is extensive and includes:

  • Analysis, archaeometry of artifacts
  • Diagnosis and monitoring for restoration and preventive conservation
  • Geosciences / Geomatics for cultural heritage
  • Education
  • Smart and sustainable tourism
  • Urban planning
  • Digital Twins


Important dates

  • Paper submission: July 19 12.59 PM, July 24, 2024 UTC-0
  • Author acceptance notification: August 5, 2024
  • Camera-Ready: August 19, 2024
  • Workshop date: TBA (28 Oct - 1 Nov, 2024)


Submission guidelines

Submission format All submissions must be original work not under review at any other workshop, conference, or journal. The workshop will accept papers describing completed work (full paper) as well as work in progress (short paper). Two submission formats are accepted: a) 4 pages plus 1-page reference (short paper); or b) 8 pages plus up to 2-page reference (full paper). They must be encoded as PDF using the ACM Article Template of the main conference ACM Multimedia 2024 (https://2024.acmmm.org/regular-papers).

Peer Review and publication in ACM Digital Library Paper submissions must conform with the “double-blind” review policy. All papers will be peer-reviewed by experts in the field, they will receive at least two reviews. Acceptance will be based on relevance to the workshop, scientific novelty, and technical quality. Depending on the number, maturity and topics of the accepted submissions, the work will be presented via oral or poster sessions. The workshop papers will be published in the ACM Digital Library.

Profile Registration Openreview (submissions' portal) requires a profile with OpenReview.

IMP NOTES:

  • New profiles created without an institutional email will go through a moderation process that can take up to two weeks.
  • New profiles created with an institutional email will be activated automatically.

Submissions' Site https://openreview.net/group?id=acmmm.org/ACMMM/2024/Workshop/SUMAC


Special Highlights

Best Paper Award - We will present a best paper award, accompanied with a certificate and a trophy, similar to previous editions.


Keynote Speakers

Jing Zhang
Dr.
Jing Zhang
School of Computer Science, The University of Sydney
Website

Dr. Jing Zhang is currently a Research Fellow at the School of Computer Science, Faculty of Engineering, University of Sydney. His research focuses on computer vision and deep learning, with over 100 CORE A* ranked publications in prestigious journals and conferences such as CVPR, ICCV, NeurIPS, ACM Multimedia, IEEE TPAMI, and IJCV. His work has garnered more than 8,000 citations according to Google Scholar. Dr. Zhang is an Area Chair for ICPR, a Senior Program Committee Member for AAAI and the IJCAI, and a guest editor for a special issue on IEEE Transactions on Big Data. He is also a regular reviewer for several leading journals and conferences. In 2023, he was promoted to IEEE Senior Member and was listed among the World’s Top 2% Scientists by Stanford University in 2022 and 2023. Dr. Zhang's groundbreaking work centers on the ViTAE series of foundation models, marking a significant advancement in vision transformers. These models incorporate convolutional biases to enhance representation capabilities and have reached an impressive 1 billion parameters. The ViTAE-H model achieved a remarkable 91.2% top-1 accuracy, outperforming industry giants like Google and Meta. He also developed ViTPose, a 1 billion-parameter pose estimation model, shows great promise for applications in human and animal behavior analysis. His contributions have significantly impacted various image-related tasks and earned widespread recognition in both academia and industry. His commitment to open-source code on GitHub has fostered collaboration, amassing over 5000 stars.

Vera Moitinho de Almeida
Dr.
Vera Moitinho de Almeida
Centre for Digital Culture and Innovation, Faculty of Arts and Humanities, University of Porto
Website

Vera Moitinho de Almeida is a senior researcher and coordinator of the Centre of Digital Culture and Innovation (CODA) at the Faculty of Arts & Humanities of the University of Porto (FLUP), a member of CITCEM-UP and INESCC-UC, and an honorary senior research collaborator at LAQU-UAB. She has an unusual interdisciplinary academic background: a PhD (cum laude) from UAB, focusing on technological and functional analysis of archaeological objects using 3D digital models and reverse engineering processes; an MSc in Prehistoric Archaeology (UAB); an interdisciplinary MSc (cum laude) in Multimedia Technologies (FEUP); and a BA in Fine Arts (IPC), with a major in pedagogy, while having attended several courses in distinct fields. Her research expertise consists of three intertwined subjects: 1) 2D/3D digital imaging and visualizations for research and conservation of cultural heritage materials; 2) Computational archaeology; 3) Digital data lifecycle and FAIR principles. She has been actively involved in several international transdisciplinary scientific projects and has published extensively in the field of digital applications to research in archaeology, cultural heritage, and the digital humanities.


Schedule (Melbourne time, 28 Oct / 1 Nov 2024)

Keynotes: 40 min talk + 10 min Q&A; Orals: 20 min talk + 5 min Q&A

Time Title Presenter
TBA (KeyNote 1) "From Pixels to Preservation - The Power of Large Vision Models in Heritage Content Understanding." Dr. Jing Zhang
(Abstract 1) "Preserving cultural heritage is essential for maintaining the legacy and history of human civilization, but it presents challenges in managing vast amounts of historical artifacts and documents. Recent advances in artificial intelligence, especially large vision models (LVMs), offer unprecedented capabilities for understanding and preserving heritage content. This keynote will explore how LVMs are revolutionizing image analysis, encompassing advancements in architecture design, multimodal learning, and image generation. We will delve into specific applications such as object detection for identifying artifacts, pose estimation for interpreting historical depictions, text detection and recognition for deciphering ancient scripts, and visual grounding for contextualizing objects within textual descriptions. By leveraging the power of LVMs, we can gain deeper insights into cultural heritage, enabling better preservation strategies and fostering a richer understanding of our past for future generations. The future holds immense promise for LVMs in this domain, with potential applications in automating restoration processes, identifying damage at an early stage, uncovering hidden stories from historical artifacts, and creating engaging, interactive experiences for wider audiences."
TBA (KeyNote 2) "God or the devil are in the details too. Reusing 3D digital resources for cultural heritage research." Dr. Vera Moitinho de Almeida
(Abstract 2) "In material cultural heritage research practice, little attention seems to have been given to the digital data lifecycle. And, although often mentioned, to making research data actually FAIR (Findable, Accessible, Interoperable, Reusable; Wilkinson et al., 2016). In this talk, I will showcase an investigation on archaeological Greek and Cypriot pottery (c. 10th-4th BCE), with a strong emphasis on 3D digital resources. The objectives are manifold, namely - to understand the chronological and geographical variability of their production, use, and shape; as well as to enable conservation studies and monitoring. However, despite the considerable amount of 3D models of archaeological objects available online and elsewhere, I will demonstrate some of the barriers encountered to reusing them and proceeding with this investigation. This presentation tackles issues related to digital repositories and archives, quality and trust of 3D digital resources, contextual and useful metadata for research (including paradata), and data incompleteness. Knowing that although a resource cannot and need not be fully described, it can be better described."


Program Committee

  • Margarita Roos (Fujitsu)
  • Marin Ferecatu (Cnam, France)
  • Prathmesh Madhu (FAU, Germany)
  • Mathias Zinnen (FAU, Germany)
  • Giulio Poggi (Cultural Heritage Technologies, IIT, Italy)
  • Milind Padalkar (PAVIS-IIT, Italy)
  • Martin Langner (Uni. Göttingen, Germany)
  • Jenny Benois-Pineau (Université de Bordeaux/LABRI, France)
  • John Samuel (CPE Lyon, France)
  • Edgar Román (ITAM, Mexico)
  • Jing Zhang (The Uni. of Sydney, Australia)
  • Stephane Marchand-Maillet (University of Geneva)
  • Ronak Gupta (Innovation Hub, IIT-Jodhpur, India)
  • Sinem Aslan (University of Venice, Italy)
  • Ziqian Luo (Oracle, USA)
  • Armanda Rodrigues (NOVA LINCS, NOVA Univ. Lisbon, Portugal)
  • Wen Wang (Zhejiang University, China)


Organizers

Valerie Gouet-Brunet
Valerie Gouet-Brunet
IGN, Gustave Eiffel University, France
Website
Ronak Kosti
Ronak Kosti
Pattern Recognition Lab, FAU, Germany
Website
Li Weng
Li Weng
Zhejiang Financial College, China
Website


Sponsors


Previous Editions