University of Trento · DISI
Multimedia
& Human
Understanding
Group
A research group at the University of Trento working on computer vision, video understanding, and multimodal AI — building models that see, listen, and reason about the world.
MHUG - Based in Trento
Research Areas
Vision-Language and Multimodal Models
Vision-language alignment with large multimodal foundation models.
3D Vision and Spatial Scene Understanding
Spatial scene reconstruction and semantic segmentation.
Geometric Deep Learning and Non-Euclidean Networks
Representation learning with non-Euclidean and geometric neural networks.
Generative AI
Image synthesis and degradation-agnostic restoration with attention mechanisms and manifold regularization.
Human-Centric Analysis and Motion Understanding
Human motion analysis and privacy-preserving vision with action-aligned representationsn.
Trustworthy AI and Adversarial Security
Trustworthy AI and adversarial defense with vision-language and multi-view systems.
Featured Projects
ELIAS
2023 – 2027
ELLIOT
2025 – 2029
GUIDANCE
2025 – 2028