Volumetric Mapping for Long-term Robot Interaction | JRC Workshop 2021

Channel:
Microsoft Research
Subscribers:
344,000
Published on ● Video Link: https://www.youtube.com/watch?v=XiHJiNu1R6s


Duration: 15:58
410 views
7

Computer Vision | Day 1
20 April 2021

Speaker: Lukas Schmid, ETH Zurich
(collaboration with Cesar Cadena, Roland Siegwart, ETH Zurich and Johannes Schönberger, Juan Nieto, Marc Pollefeys, Microsoft)

Having a high-quality map, i.e. a geometric and semantic understanding of one's surroundings over longer periods of time, is a prerequisite for a large variety of interactive tasks in robotics and AR, such as navigation, object search, manipulation, immersive AR experiences, or service robotics. However, state-of-the-art approaches are not yet well equipped to capture long-term dynamics in changing scenes. In this talk, we present current work a novel map representation leveraging semantic scene understanding to capture temporal changes in long-term dynamic environments and to provide the agent with object-centric multi-resolution maps.

Learn more about the Joint Research Center Workshop 2021: https://www.microsoft.com/en-us/research/event/joint-research-centre-workshop-2021/



Other Videos By Microsoft Research


2021-06-10HovercRaft: Achieving Scalability and Fault-tolerance for Microsecond-scale Datacenter Services
2021-06-10Noise*: A library of Verified High-Performance Secure Channel Protocol Implementations
2021-06-10Reasoning about the TLA+ operator ENABLED within TLAPS | JRC Workshop 2021
2021-06-10DataScope: Scaling up Data Shapley over Machine Learning Pipelines | JRC Workshop 2021
2021-06-09H2O: Two Hands Manipulating Objects for First Person Interaction Recognition | JRC Workshop 2021
2021-06-09Learning Motion Priors for 4D Human Body Capture in 3D Scenes | JRC Workshop 2021
2021-06-09Test-time Adaptable Neural Networks for Robust Medical Image Segmentation | JRC Workshop 2021
2021-06-09Privacy-Preserving Image Features | JRC Workshop 2021
2021-06-09Multi-view Appearance Super-resolution with CNNs | JRC Workshop 2021
2021-06-09Privacy Preserving Structure-from-Motion | JRC Workshop 2021
2021-06-09Volumetric Mapping for Long-term Robot Interaction | JRC Workshop 2021
2021-06-09Controllable Human Motion Generation from Trajectories | JRC Workshop 2021
2021-06-09Towards Markerless Surgical Tool and Hand Pose Estimation | JRC Workshop 2021
2021-06-09Project Altair: Infrared Vision and AI-Decision Making for Longer Drone Flights
2021-06-09Digital Characters in Virtual Experiences | JRC Workshop 2021
2021-06-09Reconstructing 3D Human with Learning-based Method | JRC Workshop 2021
2021-06-09Freetures: Localization in Signed Distance Function Maps | JRC Workshop 2021
2021-06-03Racist Tropes & Labor Discipline: How Tech Inherits & Reproduces Global Imaginaries of Race and Work
2021-06-02Directions in ML: Latent Stochastic Differential Equations: An Unexplored Model Class
2021-05-27Fuzzing to improve the security and reliability of cloud services with RESTler
2021-05-27Pushing the frontier of neural text to speech