Title: Video-based identity proofing

Abstract: Remote IDentity Proofing (RIDP) represents a convenient option for enabling enrolment and access to different services without requiring the physical presence of the applicant. In commonly used pipelines, the applicant presents a valid ID document and acquires a selfie through his portable or desktop device, which is then compared to the portrait photo contained in the ID document in an either manual, semi- or fully automated fashion.

The talk will provide an overview of different technical aspects and security implications of such processes, as well as of the role they can play in digital identity frameworks such as the european eIDAS ecosystem. This includes the discussion of the pros and cons implied by the use of video streams in remote environments, with focus on the advanced impersonation threats now enabled by AI-based video synthesis techniques and the different attack vectors within RIDP systems, as witnessed by recent reports.

We will then overview ongoing research efforts in formalizing dedicated threat models, as well as devising corresponding structural and detection strategies against attacks based on visual manipulation.

Bio: Cecilia Pasquini (MS’12, PhD’16) is a researcher at the Center for CybersecurityLink opens in a new window of the Fondazione Bruno KesslerLink opens in a new window (FBK), Italy, where she works with the Security and TrustLink opens in a new window research unit on media processing techniques for digital identity systems. Prior to that, she has been a researcher at the Media LabLink opens in a new window of the University of Trento (Italy), a postdoctoral fellow at the Privacy and Security LabLink opens in a new window with the University of Innsbruck (Austria) and a visiting scholar at Signal Processing in Communications Group of the University of Vigo (Spain). Her research area lies at the intersection of information security and image/video processing, with focus on media security and forensics, synthetic media detection and adversarial machine learning. She has participated in several projects on these topics funded by national (MUR, MISE, IPZS, EUREGIO) and international agencies (DARPA, EU, DFG). She has been General Co-Chair and Program Chair of the ACM Workshop on Information Hiding & Multimedia Security (in 2018 and 2023), and co-organizer of the special session "Information security meets adversarial examples" at IEEE WIFS 2019. She is an elected member of the IEEE IFS TC and of the Eurasip BForSec TAC.