Search Results for author: Paul H. J. Kelly

Found 17 papers, 6 papers with code

Gaussian Splatting SLAM

no code implementations11 Dec 2023 Hidenobu Matsuki, Riku Murai, Paul H. J. Kelly, Andrew J. Davison

We present the first application of 3D Gaussian Splatting to incremental 3D reconstruction using a single moving monocular or RGB-D camera.

3D Reconstruction Novel View Synthesis +1

Systematic Comparison of Path Planning Algorithms using PathBench

no code implementations7 Mar 2022 Hao-Ya Hsueh, Alexandru-Iosif Toma, Hussein Ali Jaafar, Edward Stow, Riku Murai, Paul H. J. Kelly, Sajad Saeedi

An unified path planning interface that facilitates the development and benchmarking of existing and new algorithms is needed.

Benchmarking

A Robot Web for Distributed Many-Device Localisation

no code implementations7 Feb 2022 Riku Murai, Joseph Ortiz, Sajad Saeedi, Paul H. J. Kelly, Andrew J. Davison

We show that a distributed network of robots or other devices which make measurements of each other can collaborate to globally localise via efficient ad-hoc peer to peer communication.

Cain: Automatic Code Generation for Simultaneous Convolutional Kernels on Focal-plane Sensor-processors

1 code implementation21 Jan 2021 Edward Stow, Riku Murai, Sajad Saeedi, Paul H. J. Kelly

Focal-plane Sensor-processors (FPSPs) are a camera technology that enable low power, high frame rate computation, making them suitable for edge computation.

Code Generation

Lossy Checkpoint Compression in Full Waveform Inversion: a case study with ZFPv0.5.5 and the Overthrust Model

no code implementations26 Sep 2020 Navjot Kukreja, Jan Hueckelheim, Mathias Louboutin, John Washbourne, Paul H. J. Kelly, Gerard J. Gorman

This paper proposes a new method that combines check-pointing methods with error-controlled lossy compression for large-scale high-performance Full-Waveform Inversion (FWI), an inverse problem commonly used in geophysical exploration.

Computational Physics Numerical Analysis Numerical Analysis

AnalogNet: Convolutional Neural Network Inference on Analog Focal Plane Sensor Processors

no code implementations2 Jun 2020 Matthew Z. Wong, Benoit Guillard, Riku Murai, Sajad Saeedi, Paul H. J. Kelly

We present a high-speed, energy-efficient Convolutional Neural Network (CNN) architecture utilising the capabilities of a unique class of devices known as analog Focal Plane Sensor Processors (FPSP), in which the sensor and the processor are embedded together on the same silicon chip.

BIT-VO: Visual Odometry at 300 FPS using Binary Features from the Focal Plane

no code implementations23 Apr 2020 Riku Murai, Sajad Saeedi, Paul H. J. Kelly

Focal-plane Sensor-processor (FPSP) is a next-generation camera technology which enables every pixel on the sensor chip to perform computation in parallel, on the focal plane where the light intensity is captured.

Visual Odometry

Scalable Uncertainty for Computer Vision with Functional Variational Inference

no code implementations CVPR 2020 Eduardo D. C. Carvalho, Ronald Clark, Andrea Nicastro, Paul H. J. Kelly

As Deep Learning continues to yield successful applications in Computer Vision, the ability to quantify all forms of uncertainty is a paramount requirement for its safe and reliable deployment in the real-world.

Depth Estimation Gaussian Processes +3

Architecture and performance of Devito, a system for automated stencil computation

3 code implementations9 Jul 2018 Fabio Luporini, Michael Lange, Mathias Louboutin, Navjot Kukreja, Jan Hückelheim, Charles Yount, Philipp Witte, Paul H. J. Kelly, Gerard J. Gorman, Felix J. Herrmann

Some of these are obtained through well-established stencil optimizers, integrated in the back-end of the Devito compiler.

Mathematical Software 65N06, 68N20

Algorithmic Performance-Accuracy Trade-off in 3D Vision Applications Using HyperMapper

no code implementations2 Feb 2017 Luigi Nardi, Bruno Bodin, Sajad Saeedi, Emanuele Vespa, Andrew J. Davison, Paul H. J. Kelly

In this paper we investigate an emerging application, 3D scene understanding, likely to be significant in the mobile space in the near future.

Active Learning Scene Understanding

Comparative Design Space Exploration of Dense and Semi-Dense SLAM

no code implementations15 Sep 2015 M. Zeeshan Zia, Luigi Nardi, Andrew Jack, Emanuele Vespa, Bruno Bodin, Paul H. J. Kelly, Andrew J. Davison

SLAM has matured significantly over the past few years, and is beginning to appear in serious commercial products.

Benchmarking

Firedrake: automating the finite element method by composing abstractions

1 code implementation8 Jan 2015 Florian Rathgeber, David A. Ham, Lawrence Mitchell, Michael Lange, Fabio Luporini, Andrew T. T. McRae, Gheorghe-Teodor Bercea, Graham R. Markall, Paul H. J. Kelly

Firedrake is a new tool for automating the numerical solution of partial differential equations.

Mathematical Software Numerical Analysis Numerical Analysis G.1.8; G.4

Introducing SLAMBench, a performance and accuracy benchmarking methodology for SLAM

3 code implementations8 Oct 2014 Luigi Nardi, Bruno Bodin, M. Zeeshan Zia, John Mawer, Andy Nisbet, Paul H. J. Kelly, Andrew J. Davison, Mikel Luján, Michael F. P. O'Boyle, Graham Riley, Nigel Topham, Steve Furber

Real-time dense computer vision and SLAM offer great potential for a new level of scene modelling, tracking and real environmental interaction for many types of robot, but their high computational requirements mean that use on mass market embedded platforms is challenging.

Benchmarking

A thread-parallel algorithm for anisotropic mesh adaptation

2 code implementations12 Aug 2013 Georgios Rokos, Gerard J. Gorman, James Southern, Paul H. J. Kelly

Inter-node parallelism for mesh adaptivity has been successfully implemented by a number of groups.

Distributed, Parallel, and Cluster Computing

SLAM++: Simultaneous Localisation and Mapping at the Level of Objects

no code implementations CVPR 2013 Renato F. Salas-Moreno, Richard A. Newcombe, Hauke Strasdat, Paul H. J. Kelly, Andrew J. Davison

We present the major advantages of a new 'object oriented' 3D SLAM paradigm, which takes full advantage in the loop of prior knowledge that many scenes consist of repeated, domain-specific objects and structures.

3D Object Recognition Descriptive +2

Cannot find the paper you are looking for? You can Submit a new open access paper.