Toward Relational Virtual Humans: capturing, matching, and hallucinating 3D interactions

Relatore:  Riccardo Marin - Università di Tubinga (Ger
  martedì 27 giugno 2023 alle ore 11.30 Sala Verde (solo in presenza)
Abstract: The industry has a substantial bet on extended reality (XR). Blending the digital and real worlds in a unified interactive environment promises a new way to connect with each other, empowering and augmenting our relations. This ambitious plan requires user-centric technology, boosting the research in Virtual Humans. Acquiring, relating, and synthesizing 3D human identities and poses are long-standing (and yet-to-be-solved) problems. But a new set of challenges need to be solved: users wear disparate garments, significantly interact with items in the surrounding environments, and even use hand gestures to manipulate digital objects that lack a physical counterpart. Furthermore, XR applications start from in-the-wild data, like measurements from sparse wearables (e.g., IMUs) or noisy sensors (e.g., depth scans). In this talk, I will present recent works about capturing interactions in large scenes, human registration, and object hallucination from 3D human point clouds, pointing to exciting new challenges and providing practical tools to foster research in XR.
 
Bio: 
Riccardo is a Post Doctoral researcher at the University of Tuebingen in the Real-Virtual Humans group led by Gerard Pons-Moll, funded by a Humboldt Foundation Post-Doctoral Fellowship and a Marie Sklodowska-Curie Individual Fellowship, and a Member of the ELLIS Society. Previously, he was a post-doc at the Sapienza University of Rome in the GLADIA group led by Emanuele Rodolà, leading a work package of the SPECGEO ERC project. Riccardo followed the University of Verona’s Ph.D. program in Computer Science under the supervision of Umberto Castellani, collecting a Best PhD Thesis award from the Italian Chapter of EuroGraphics (EG-Italy). His research focuses on Spectral Shape Analysis, Shape Matching, Geometric Deep Learning, and Virtual Humans. His work appears in top-level conferences and journals of machine learning, computer vision and graphics (NeurIPS, CVPR, IJCV, CGF), obtaining a Best Student Paper Award (3DV 2020). He often serves as a lecturer for tutorials and doctoral schools on geometry processing topics at international conferences, as a conference organizer, DEI-Iniative promoter, and reviewer, for which he collected five Outstanding Reviewer Awards.
 
Referente: Umberto Castellani

Referente
Umberto Castellani

Referente esterno
Data pubblicazione
19 giugno 2023

Offerta formativa

Condividi