On June 12th, the Computer Vision Trento Symposium 2024 will gather experts and present new research in computer vision. This year’s schedule includes a variety of topics exploring recent advancements in the field. The event is organized by the Fondazione Bruno Kessler, our project partner, and the University of Trento. We are proud to be recognized as the event’s partner.
This one-day event will showcase a series of oral and poster presentations, including 12 works that will be presented at the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2024, the most important conference in its field.
Presentations will cover subjects such as object and scene understanding, 3D reconstruction, image editing, and bias detection from a variety of input data representations, including 2D images, 3D point clouds, and videos. They will explore recent advances in vision foundation models, large language models, vision-language models, and diffusion models.
We would like to highlight a particular piece of research that was supported by AI-PRISM, titled “Open-vocabulary object 6D pose estimation”, which will be presented by Jaime Corsetti (the first author of the paper co-authored by Davide Boscaini, Changjae Oh, Andrea Cavallaro, and Fabio Poiesi). This paper was accepted at CVPR 2024 as a highlight poster (with an acceptance rate of 2.8%). It enables the prediction of an object’s rotation and translation within a scene based solely on the object’s textual description. This setting is particularly relevant for AI-PRISM because it enhances robotic perception capabilities, enables seamless robotic interaction with objects, and allows for intuitive natural language interaction with human workers.