Building digital twins help bridge the physical world we live in along with the digital one in a large scale, persistent manner within a community of people. This tutorial provides an overview of the computer vision and deep learning work that is behind making large scale digital twins a reality with a focus on enabling two key aspects: Persistence and Large Scale. The four fundamental pillars of Perception to enable this functionality are introduced: Environment mapping, persistence, understanding (in terms of objects and scenes) and scalability. The tutorial will provide a deep dive into the main building blocks along with key challenges and open problems.

Check out our program here.

Check out our CVPR 2019 tutorial site here.