Getting Robots In The Future To Truly See

Getting Robots In The Future To Truly See

Developing robots that can process visual information in real-time could lead to a new range of handy and helpful robots for around the home and in industry. Professor Andrew Davison and Dr Stefan Leutenegger from the Dyson Robotics Lab at Imperial College London discuss the advances they are making in developing robotic vision.

2021 Widget

CodeMapping

CodeMapping

CodeMapping

CodeMapping: Real-Time Dense Mapping for Sparse SLAM using Compact Scene Representations

End-to-End Egospheric Spatial Memory

End-to-End Egospheric Spatial Memory

End-to-End Egospheric Spatial Memory

iMAP

iMAP

iMAP: Implicit Mapping and Positioning in Real-Time

In-Place Scene Labelling

In-Place Scene Labelling

In-Place Scene Labelling and Understanding with Implicit Scene Representation

NodeSLAM

NodeSLAM

NodeSLAM: Neural Object Descriptors for Multi-View Shape Reconstruction

SIMstack

SIMstack

SIMstack: A Generative Shape and Instance Model for Unordered Object Stacks

2020 Widget

MoreFusion

MoreFusion

MoreFusion

Comparing View-Based and Map-Based Semantic Labelling in Real-Time SLAM

DeepFactors

DeepFactors

DeepFactors: Real-Time Probabilistic Dense Monocular SLAM

Learning One-shot Imitation

Learning One-shot Imitation

Learning One-Shot Imitation from Humans without Humans

MoreFusion

MoreFusion

MoreFusion: Multi-object Reasoning for 6D Pose Estimation from Volumetric Fusion

RLBench

RLBench

RLBench: The Robot Learning Benchmark

2019 Widget

KO Fusion

KO Fusion

KO Fusion

KO Fusion: Dense Visual SLAM with TIghtly-Coupled Kinematic and Odometric Tracking

Learning Meshes for Dense Visual SLAM

Learning Meshes for Dense Visual SLAM

Learning Meshes for Dense Visual SLAM

SceneCode

SceneCode

SceneCode: Monocular Dense Semantic Reconstruction using Learned Encoded Scene Representations

X-Section

X-Section

X-Section: Cross-Section Prediction for Enhanced RGB-D Fusion

2018 Widget

CodeSLAM

CodeSLAM

CodeSLAM

CodeSLAM - Learning a Compact, Optimisable Representation for Dense Visual SLAM.

Few-Shot Imitation Learning

Few-Shot Imitation Learning

Task-Embedded Control Networks for Few-Shot Imitation Learning

Fusion++

Fusion++

Fusion++: Volumetric Object-Level SLAM

LS-Net

LS-Net

LS-Net: Learning to Solve Nonlinear Least Squares for Dense Tracking and Mapping

2017 Widget

Dense RGB-D-Inertial SLAM with Map Deformations

Dense RGB-D-Inertial SLAM with Map Deformations

Dense RGB-D-Inertial SLAM with Map Deformations

Dense RGB-D-Inertial SLAM with Map Deformations

Multi-Stage Task

Multi-Stage Task

Transferring End-to-End Visuomotor Control from Simulation to Real World for a Multi-Stage Task

Semantic Texture for Robust Dense Tracking

Semantic Texture for Robust Dense Tracking

Semantic Texture for Robust Dense Tracking

SemanticFusion

SemanticFusion

SemanticFusion: Dense 3D Semantic Mapping with Convolutional Neural Networks

2016 Widget

Deep Learning a Grasp Function

Deep Learning a Grasp Function

Deep Learning a Grasp Function

Deep Learning a Grasp Function for Grasping under Gripper Pose Uncertainty

Event Camera

Event Camera

Simultaneous Optical Flow and Intensity Estimation from an Event Camera

Height Map Fusion

Height Map Fusion

Real Time Height Map Fusion using Differentiable Rendering

Monocular, Real-Time Surface Reconstruction

Monocular, Real-Time Surface Reconstruction

Monocular, Real-Time Surface Reconstruction using Dynamic Level of Detail

Elastic Fusion Widget

ElasticFusion: Dense SLAM Without A Pose Graph

The video above demonstrates the Elastic Fusion system, a novel approach to real-time dense visual SLAM. Our approach applies local model-to-model surface loop closure optimisations to stay close to the mode of the map distribution, while utilising global loop optimisations to recover from arbitrary drift and maintain global consistency.

T Whelan, S Leutenegger, B Glocker, R F. Salas-Moreno,  AJ Davison, ElasticFusion: Dense SLAM Without A Pose Graph. Robotics: Science and Systems (RSS), Rome, Italy, July 2015

ElasticFusion: Dense SLAM Without A Pose Graph

ElasticFusion: Dense SLAM Without A Pose Graph

Demostration of real time ElasticFusion on an office, hotel and copy dataset

The video above demonstrates the Elastic Fusion system, a novel approach to real-time dense visual SLAM. Our approach applies local model-to-model surface loop closure optimisations to stay close to the mode of the map distribution, while utilising global loop optimisations to recover from arbitrary drift and maintain global consistency.

T Whelan, S Leutenegger, B Glocker, R F. Salas-Moreno,  AJ Davison, ElasticFusion: Dense SLAM Without A Pose Graph. Robotics: Science and Systems (RSS), Rome, Italy, July 2015

ElasticFusion: Dense SLAM Without A Pose Graph (extras)

ElasticFusion: Dense SLAM Without A Pose Graph (extras)

ElasticFusion on seating area, garden, The Burghers of Calais, stairs, MIT-76-417b, loopback dataset

The video above demonstrates the Elastic Fusion system, a novel approach to real-time dense visual SLAM. Our approach applies local model-to-model surface loop closure optimisations to stay close to the mode of the map distribution, while utilising global loop optimisations to recover from arbitrary drift and maintain global consistency.

T Whelan, S Leutenegger, B Glocker, R F. Salas-Moreno,  AJ Davison, ElasticFusion: Dense SLAM Without A Pose GraphRobotics: Science and Systems (RSS), Rome, Italy, July 2015