Oral presentation (15 min)IJCB 2026 Focus Session on Generative AI for Fair and Secure Biometrics under Limited Data, Rome, Italy
Fifteen-minute oral presentation in the Focus Session on Generative AI for Fair and Secure Biometrics under Limited Data, plus a poster.
Oral presentation (15 min)IEEE International Joint Conference on Biometrics (IJCB) 2026, Rome, Italy
Fifteen-minute oral presentation in the main conference.
Poster spotlight (2 min) and posterIEEE International Joint Conference on Biometrics (IJCB) 2026, Rome, Italy
Two-minute poster spotlight in the main conference, followed by the poster session.
Poster spotlight (2 min) and posterIEEE International Joint Conference on Biometrics (IJCB) 2025, Osaka, Japan
Two-minute poster spotlight in the main conference, followed by the poster session.
Poster and demoMartigny Biometrics Workshop 2025 (EAB-CITeR), Martigny, Switzerland
Poster and live demo at the 2025 Martigny Biometrics Workshop, organized with the European Association for Biometrics (EAB) and CITeR. The demo showed our method for aging and de-aging faces by editing the StyleGAN2 latent space while keeping the identity of the person, ahead of its publication at IJCB 2025.
Virtual presentation (video) and posterInternational Conference on Pattern Recognition (ICPR) 2024, Kolkata, India (presented remotely)
SwiftFaceFormer is an efficient and lightweight hybrid architecture that combines convolutional and transformer components for accurate face recognition on resource-constrained devices. I presented the paper virtually with a recorded video; the poster was displayed at the conference venue.
TalkForum InCyber (FIC) 2024, presenting the EU H2020 SOTERIA project, Lille, France
Talk on behalf of the EU H2020 SOTERIA project at Forum InCyber (FIC) 2024 in Lille. SOTERIA developed a user-centric tool for European citizens to manage their personal data in a decentralized, secure data vault. The talk presented our work at Inria on bringing machine learning to the data vault while preserving privacy, using decentralized learning and Trusted Execution Environments.
PosterLatinX in AI (LXAI) Workshop at ICCV 2023, Paris, France
Poster presentation of our peer-reviewed extended abstract at the LatinX in AI workshop at ICCV 2023. The work explores whether binarized networks can bring face recognition close to Vision Transformer-level accuracy while running in real time on embedded systems, continuing my PhD research on efficient face recognition.
Doctoral consortium presentationCIDTEC Doctoral Consortium, Tecnológico de Monterrey (virtual edition), Online
I presented my PhD research on affordable, real-time very low resolution face recognition for automated video surveillance. In surveillance footage, faces are often only a few pixels wide, and state-of-the-art recognition models are too heavy to run on the embedded devices where cameras are deployed. My thesis addressed both problems with lightweight and binarized neural networks for very low resolution face recognition.