Computer Vision: the Deep Learning Power and the Geometric Wisdom
DEIB - Conference Room "E. Gatti" (Bldg. 20)
December 6th, 2022
5.30 pm
Contacts:
Nicola Gatti
Research Line:
Artificial intelligence and robotics
December 6th, 2022
5.30 pm
Contacts:
Nicola Gatti
Research Line:
Artificial intelligence and robotics
Abstract
In the framework of the AI Seminars 2022, on December 6th, 2022 at 5.30 pm, the seminar titled "Computer Vision: the Deep Learning Power and the Geometric Wisdom" will be held by Giacomo Boracchi, Federica Arrigoni and Luca Magri in DEIB Conference Room.
Deep Learning has revolutionized Computer Vision, successfully addressing visual recognition problems that were considered impossible to solve a few years ago. Despite such a dazzling success, a huge number of problems in 3D reconstruction and imaging are still solved by accurate and elegant algorithms relying on geometry and explicit modeling, as opposed to implicitly learning from data. In this talk, we will illustrate our research projects in both visual recognition, nowadays dominated by deep learning power, and 3D Vision, where Geometric wisdom still plays a primary role.
The registration is available at the following LINK.
For organizational reasons, only the first 100 registered participants will be admitted.
Those who register later will be sent a link to follow the event in streaming
More information in attachment.
Deep Learning has revolutionized Computer Vision, successfully addressing visual recognition problems that were considered impossible to solve a few years ago. Despite such a dazzling success, a huge number of problems in 3D reconstruction and imaging are still solved by accurate and elegant algorithms relying on geometry and explicit modeling, as opposed to implicitly learning from data. In this talk, we will illustrate our research projects in both visual recognition, nowadays dominated by deep learning power, and 3D Vision, where Geometric wisdom still plays a primary role.
The registration is available at the following LINK.
For organizational reasons, only the first 100 registered participants will be admitted.
Those who register later will be sent a link to follow the event in streaming
More information in attachment.