Several things happened since my last update, which has been a while ago. We published MaskFusion at ISMAR2018, I interned and contracted for Facebook Reality Labs in Redmond, WA and in the course of that we are going to publish FroDO at CVPR2020 – which sadly has to be purely virtual.
The two publications follow my trail of work on modelling scenes in an object-aware fashion. While MaskFusion is a real-time method for the tracking and mapping of dynamic objects from RGBD video, FroDO is object-aware on the reconstruction level and facilitates a shape embedding which functions as a shape prior.
The time in the US has been great, working with a fantastic team in an amazing environment. Just look at how beautiful Mnt. Rainier is :