Menu Close

Special Sessions

1 – Forensics and Security of Physical Objects

Iuliia Tkachenko (LIRIS, CNRS, Université Lumière Lyon 2, France)
Justin Picard (Scantrust)
Slava Voloshynovskiy (University of Geneva, Suisse)

Short description
Globalization and improvements in digital scanning and printing technologies have made counterfeiting more prolific and easier to perform than ever. According to a report commissioned by the International Chamber of Commerce, the entire global economy is on track to lose €3.7 trillion to counterfeiting and piracy with 5.4 million jobs at risk by 2022. Counterfeiting also has a significant impact on our health and safety. According to the World Health Organization, up to half of malaria medications could be fake, for a disease that kills around one million people globally.
Counterfeiting and forgery continues to proliferate partly due to the limitations of existing anti-counterfeiting technologies. Most of these technologies are either too easy to copy, too expensive to implement, too cumbersome to authenticate, or some combination therein. It has become trivial for counterfeiters to make visually perfect copies of physical products and the majority of the more secure methods for verifying authenticity cannot be used by consumers.
This special session aims at covering both forensics techniques (printer/scanner identification, 3D printer fingerprints, paper and ink identification) and security techniques (copy sensitive codes, natural and artificial randomness, security printing, printable digital watermarks, PUFs, etc.) for protection of physical objects, such as packaging, branded products, security documents, against counterfeiting, forgery, tampering and integrity attacks. Related security techniques, such as techniques for the protection of 3D models or IoT techniques for the communication between physical objects are also welcome.
The session will focus on finding the links between forensics, image and signal processing, machine learning and cryptography and on presentation of the novel techniques to digitally assess the authenticity and integrity of physical items. The session encourages contributions focused on novel techniques for physical object security, on making the existing protection solutions accessible to consumers on a large scale (e.g. by smartphone), and on the study of the security limits of existing techniques.

Topics of interest
  • Physical object protection, authentication and identification
  • Physical object integrity check
  • Printer/scanner fingerprints
  • Natural and artificial randomness
  • Digital techniques for security printing
  • Paper/ink identification
  • Printable digital watermarks
  • Physical unclonable functions for IoT and object protection

  • 2 – Security for Healthcare Applications

    Victor Sanchez (University of Warwick, UK)

    Short description
    The increasing use of e-healthcare technologies has resulted in large amounts of multimedia data, i.e., images, video, audio, and vital signs, that are now an integral part of medical environments. For example, medical images and videos are now key pieces of information in electronic patient records to discover disease patterns, provide a timely diagnosis, and develop accurate treatments. Similarly, wireless sensor networks (WSNs) are now widely used to collect vital signs to monitor patients in real-time. Given the sensitive nature of these multimedia data, its security and confidentiality are of utmost importance. This is particularly important in the development of computer-assisted diagnosis (CAD) methods that rely on deep/machine learning, as a large amount of data are often needed to train the models. Security and confidentiality are also important issues in the multimedia data acquired by smart wearable medical devices, which are often connected to cloud services. This special session aims to bring together the current research progress, from both academia and industry, on security for healthcare applications. It will attract healthcare practitioners, who have access to interesting multimedia data sources and healthcare systems whose security and privacy may need to be guaranteed effectively, and researchers with security expertise.

    Topics of interest
  • Privacy-preserving deep/ machine learning for computer-assisted diagnosis (CAD) and data analysis
  • Watermarking for medical images and videos
  • Cryptography algorithms for telemedicine and cloud-based services in healthcare
  • Blockchain technologies for secure sharing healthcare datasets
  • Privacy-preservation approaches for data collected by wireless body area networks (WBANs) and wireless sensor networks (WSNs)
  • Biometric-based remote patient authentication schemes for healthcare systems
  • De-identifying techniques for multimedia data in healthcare
  • Deep/machine learning methods to develop anonymization techniques for multimedia data in healthcare

  • 3 – 3D Security

    Adrian G. Bors (Univ. of York, UK)
    Ioannis Ivrissimtzis (Univ. of Durham, UK)

    Short description
    Due to their popularity in areas related to Computer Vision and Computer Graphics, the usage of 3D graphical objects is growing continuously. Representations of 3D graphical objects, such as meshes, video, 3D point clouds, or voxel-based volumetric representations, have found numerous scientific, industrial and commercial applications, most notably in the domains of manufacturing, healthcare, video games, entertainment and creative arts, among others. In specific markets, it is also quite common for the 3D content to be commercialized by its creators, especially by those working at the intersection between the creative sector of the economy and technology, e.g. in applications of virtual reality, augmented reality, computer games, medical and scientific visualizations, cinematography. The growing popularity brings also the need to ensure the security of 3D graphical objects and to protect their databases from possible attacks. 3D watermarking has already been shown to be quite effective in authentication and copyright protection of 3D objects. On the other hand, 3D secret sharing remains an area relatively unexplored. Moreover, the increasing usage of deep learning algorithms in processing 3D objects brings new challenges on its own. For example, 3D objects used for training deep learning models may be maliciously altered, or their database be poisoned by an attacker with false objects, in order to mislead the training of a deep learning system. IEEE WIFS 2021 aims to bring together researchers from academia, industry and the creative technologies in order to discuss emerging challenges related to 3D security and how can these be tackled.

    Topics of interest
  • 3D object watermarking and steganography
  • Steganalysis of 3D objects
  • 3D object cryptography
  • Video watermarking, steganalysis and security
  • Poisoning attacks of 3D object databases
  • Backdoor attacks in 3D point clouds
  • Deep learning methods for 3D security
  • 3D security for large databases
  • Ethical and legal issues of 3D security