Bridging machine learning and computer vision for robust scene understanding

Speaker:  Alessio Del Bue - IIT Central Research Lab Genova
  Thursday, June 7, 2018 at 4:30 PM Aula I
In this talk I will present how semantic information, as
extracted by machine learning approaches, can be used to perform
classical 3D Computer Vision tasks in a robust way. In particular,
object priors can be used to simply calibrate camera motion and to
identify the position and orientation of objects in 3D. Strikingly, this
solution has a closed form that can be solved very efficiently.
Moreover, the objects position can be used to initialise efficiently
standard methods for 3D reconstruction and for higher order reasoning on
the 3D scene. Practical applications show that this computational
framework can efficiently inject semantic information in Structure from
Motion problems.

Contact person: Marco Cristani

Programme Director

Publication date
May 22, 2018