MEDIA LAB

We presented 2 papers at the ACM Multimedia Systems Conference. : “A Modular System for Enhanced Robustness of Multimedia Understanding Networks via Deep Parametric Estimation” (research track) and “Questset: A VR Dataset for Network and Quality of Experience Studies” (software and dataset track) The first work has been possible thanks to a collaboration between the MEDIA Lab and Samsung Research… Read more: 2 Papers at ACM MMSys 2024

During the GTTI-SPS Thematic Meeting on Multimedia Signal Processing the demo presented by Sara Baldoni (University of Padova) and Anna Ferrarotti (Roma Tre University) won the Best Demo Award. The 2024 edition of the GTTI-SPS Thematic Meeting on Multimedia Signal Processing, organized by the University of Trento,  was held in Andalo (TN) on January 22 and 23. The demo “Quality of… Read more: Best demo award @ GTTI MMSP 2024

HePIC dataset released with the publication of our ICIP paper in collaboration with ICEA Dept. @Unipd! We are thrilled to announce that our Heritage Pointcloud Instance Collection (HePIC) dataset has been published! It consists of a dataset of Heritage Buildings Point clouds labeled at Semantic and Instance Level, acquired via Terrestrial Laser Scanner by our colleagues of the Dept. of… Read more: HePIC dataset released

We are happy to announce that on Monday September 4 @ 14:00 in room Ke we are going to host a talk by prof. Anderson Rocha from Universidade Estadual de Campinas, Brasil titled Artificial Intelligence and Digital Forensics: a much-needed alliance Abstract: In this talk, we will discuss a panoramic view of digital forensics in the last 10 years and… Read more: Short course “Elementary, my dear human!”

We are happy to announce that the paper “Exploiting Multiple Priors for Neural 3D Indoor Reconstruction” has been accepted as oral presentation at the 34th British Machine Vision Conference (BMVC) held in Aberdeen, Scotland. This work has been possible thanks to a collaboration between the M3Dia Lab and the Sony Europe R&D department. Congratulations to the authors Federico Lincetto, Gianluca… Read more: Paper on NeRF-based 3D reconstruction accepted at BMVC 2023

The SYNDRONE dataset has been released! It is a multimodal synthetic dataset containing images and 3D data taken at multiple flying heights. It has both object-level and pixel-level annotations. It contains 72k labeled samples that allow for effective training of deep learning architectures. Download the dataset here