Perceiving Humans in the 3D World with Semantic Keypoints.
Deep Learning endeavors to make the transition from advanced driver-assist systems to fully self-driving cars. Yet the current state of technology still leads to fatal accidents and is by no means conveying trust. In this talk we present a deep learning frameworks that can accurately perceive the world in 3D by relying only on monocular or stereo cameras. We propose to escape the pixel domain using semantic keypoints, a sparse representation for every object in the scene. Driven by the limitation of neural networks outputting point estimates, we study how to estimate confidence interval for each prediction. In particular, we put emphasis on vulnerable road users, such as pedestrians and cyclist, and explicitly address the long tail of perception, to contribute to the safety of our roads.
About Lorenzo Bertoni
Lorenzo Bertoni is a doctoral student at the Visual Intelligence for Transportation (VITA) lab at EPFL in Switzerland focusing on 3D vision for vulnerable road users. Before joining EPFL, Lorenzo was a management consultant at Oliver Wyman and a visiting researcher at the University of California, Berkeley, working on predictive control for autonomous vehicles. Lorenzo received Bachelors and Masters Degrees in Engineering from the Polytechnic University of Turin and the University of Illinois at Chicago.
Lorenzo Research Page: https://www.cs.ubc.ca/~schmidtm/
The seminar will be held online via Zoom on May 7th at 14.30h CET.
To attend the seminar please send an email to [email protected]