Projects
Projects
![source_code](https://pxl-imperialacuk.terminalfour.net/fit-in/687x378/prod01/channel_2/media/migration/research-groups/source_code2--tojpeg_1483615918283_x4.jpg)
Software
CodeSLAM
A system that generates large scale photorealistic rendering of indoor scene trajectories.
DeepFactors
A real-time dense visual SLAM system capable of capturing comprehensive dense keyframe maps of room scale environments explored using an RGB camera.
ElasticFusion
A real-time dense visual SLAM system capable of capturing comprehensive dense globally consistent surfel-based maps of room scale environments explored using an RGB-D camera.
MoreFusion
A real-time robotics application where a robot arm precisely and orderly disassembles complicated piles of objects, using only on-board RGB-D vision.
ReCo
A contrastive learning framework designed at a regional level to assist learning in semantic segmentation.
SceneNet RGB-D
A system that generates large scale photorealistic rendering of indoor scene trajectories.
SemanticFusion
A real-time visual SLAM system capable of semantically annotating a dense 3D scene using Convolutional Neural Networks.
X-Section
An RGB-D 3D reconstruction approach that leverages deep learning to make object-level predic- tions about thicknesses that can be readily integrated into a volumetric multi-view fusion process.
![dataset](https://pxl-imperialacuk.terminalfour.net/fit-in/687x378/prod01/channel_2/media/migration/research-groups/dataset--tojpeg_1483616017793_x4.jpg)
Datasets
RLBench Dataset
RLBench features 400 variations of 100 completely unique, hand designed tasks ranging in difficulty, from simple target, such as reaching and door opening, to longer multi-stage tasks, such as opening an oven and placing a tray in it. The scale and diversity of RLBench offers unparalleled research opportunities in the robot learning community and beyond.
SceneNet RGB-D
Large scale photorealistic rendering of indoor scene trajectories. Random sampling permits virtually unlimited scene configurations, and here we provide a set of 5M rendered RGB-D images from over 15K trajectories in synthetic layouts with random but physically simulated object poses. Each layout also has random lighting, camera trajectories, and textures.