no code implementations • 14 Mar 2024 • Benjamin Ramtoula, Daniele De Martini, Matthew Gadd, Paul Newman
Two parallel lines of work on VPR have shown, on one side, that general-purpose off-the-shelf feature representations can provide robustness to domain shifts, and, on the other, that fused information from sequences of images improves performance.
no code implementations • 7 Mar 2024 • Georgi Pramatarov, Matthew Gadd, Paul Newman, Daniele De Martini
This paper is about 3D pose estimation on LiDAR scans with extremely minimal storage requirements to enable scalable mapping and localisation.
no code implementations • 27 Feb 2024 • David S. W. Williams, Daniele De Martini, Matthew Gadd, Paul Newman
Knowing when a trained segmentation model is encountering data that is different to its training data is important.
no code implementations • 27 Feb 2024 • David S. W. Williams, Matthew Gadd, Paul Newman, Daniele De Martini
This work proposes a semantic segmentation network that produces high-quality uncertainty estimates in a single forward pass.
no code implementations • 16 Feb 2024 • Jianhao Yuan, Shuyang Sun, Daniel Omeiza, Bo Zhao, Paul Newman, Lars Kunze, Matthew Gadd
Recent advancements in Multi-Modal Large Language models (MLLMs) have shown promising potential in enhancing the explainability as a driving agent by producing control predictions along with natural language explanations.
1 code implementation • 27 Jan 2024 • Matthew Gadd, Paul Newman
We achieve a median of 91. 52% in Recall@1, outstripping the 69. 55% for the only other open implementation, RaPlace, and at a fraction of its computational cost (relying on fewer integral transforms e. g. Radon, Fourier, and inverse Fourier).
1 code implementation • 20 Oct 2023 • Matthew Gadd, Benjamin Ramtoula, Daniele De Martini, Paul Newman
In the case of localisation, important dataset differences impacting performance are modes of appearance change, including weather, lighting, and season.
no code implementations • CVPR 2023 • Benjamin Ramtoula, Matthew Gadd, Paul Newman, Daniele De Martini
For this, we propose representing images - and by extension datasets - using Distributions of Neuron Activations (DNAs).
no code implementations • 30 Jun 2022 • Georgi Pramatarov, Daniele De Martini, Matthew Gadd, Paul Newman
This paper is about extremely robust and lightweight localisation using LiDAR point clouds based on instance segmentation and graph matching.
no code implementations • 7 Mar 2022 • Valentina Musat, Daniele De Martini, Matthew Gadd, Paul Newman
In this paper we present a compositing image synthesis method that generates RGB canvases with well aligned segmentation maps and sparse depth maps, coupled with an in-painting network that transforms the RGB canvases into high quality RGB images and the sparse depth maps into pixel-wise dense depth maps.
no code implementations • 6 Oct 2021 • Matthew Gadd, Daniele De Martini, Paul Newman
We learn, in an unsupervised way, an embedding from sequences of radar images that is suitable for solving the place recognition problem with complex radar data.
no code implementations • 16 Jun 2021 • Tarlan Suleymanov, Matthew Gadd, Daniele De Martini, Paul Newman
In this paper we present the Oxford Road Boundaries Dataset, designed for training and testing machine-learning-based road-boundary detection and inference approaches.
no code implementations • 12 Jun 2021 • Matthew Gadd, Daniele De Martini, Paul Newman
We learn, in an unsupervised way, an embedding from sequences of radar images that is suitable for solving place recognition problem using complex radar data.
no code implementations • 1 Mar 2021 • David Williams, Matthew Gadd, Daniele De Martini, Paul Newman
In this work, we train a network to simultaneously perform segmentation and pixel-wise Out-of-Distribution (OoD) detection, such that the segmentation of unknown regions of scenes can be rejected.
no code implementations • 3 Jun 2020 • Tim Y. Tang, Daniele De Martini, Shangzhe Wu, Paul Newman
Publicly available satellite imagery can be an ubiquitous, cheap, and powerful tool for vehicle localisation when a prior sensor map is unavailable.
no code implementations • 11 May 2020 • David Williams, Daniele De Martini, Matthew Gadd, Letizia Marchegiani, Paul Newman
Reliable outdoor deployment of mobile robots requires the robust identification of permissible driving routes in a given environment.
no code implementations • 2 Apr 2020 • Prannay Kaul, Daniele De Martini, Matthew Gadd, Paul Newman
This paper presents an efficient annotation procedure and an application thereof to end-to-end, rich semantic segmentation of the sensed environment using FMCW scanning radar.
no code implementations • 10 Mar 2020 • Horia Porav, Valentina-Nicoleta Musat, Tom Bruls, Paul Newman
Acquisition of data with adverse conditions in robotics is a cumbersome task due to the difficulty in guaranteeing proper ground truth and synchronising with desired weather conditions.
no code implementations • 24 Feb 2020 • Will Maddern, Geoffrey Pascoe, Matthew Gadd, Dan Barnes, Brian Yeomans, Paul Newman
We describe the release of reference data towards a challenging long-term localisation and mapping benchmark based on the large-scale Oxford RobotCar Dataset.
no code implementations • 22 Jan 2020 • Ştefan Săftescu, Paul Newman
We make the observation that features in the feature maps are viewpoint-dependent, and propose a method for transforming features with dynamic filters generated by a multi-layer perceptron from the relative poses between views.
no code implementations • 9 Jan 2020 • Tim Y. Tang, Daniele De Martini, Dan Barnes, Paul Newman
This paper is about localising a vehicle in an overhead image using FMCW radar mounted on a ground vehicle.
no code implementations • 8 Sep 2019 • Ştefan Săftescu, Paul Newman
In contrast to previous work, we pay attention to the problem of excessive smoothing in corrected meshes.
3 code implementations • 3 Sep 2019 • Dan Barnes, Matthew Gadd, Paul Murcutt, Paul Newman, Ingmar Posner
In this paper we present The Oxford Radar RobotCar Dataset, a new dataset for researching scene understanding using Millimetre-Wave FMCW scanning radar data.
Robotics Signal Processing
no code implementations • 25 Jul 2019 • Horia Porav, Tom Bruls, Paul Newman
Modern models that perform system-critical tasks such as segmentation and localization exhibit good performance and robustness under ideal conditions (i. e. daytime, overcast) but performance degrades quickly and often catastrophically when input conditions change.
no code implementations • 11 Jul 2019 • Tarlan Suleymanov, Lars Kunze, Paul Newman
Hence, we believe that our LIDAR-based approach provides an efficient and effective way to detect visible and occluded curbs around the vehicles in challenging driving scenarios.
no code implementations • 10 Jul 2019 • Tom Bruls, Horia Porav, Lars Kunze, Paul Newman
Road markings provide guidance to traffic participants and enforce safe driving behaviour, understanding their semantic meaning is therefore paramount in (automated) driving.
no code implementations • 17 May 2019 • Simon Chadwick, Paul Newman
The availability of a large quantity of labelled training data is crucial for the training of modern object detectors.
no code implementations • 25 Apr 2019 • Sarah H. Cen, Paul Newman
Radar detects stable, long-range objects under variable weather and lighting conditions, making it a reliable and versatile sensor well suited for ego-motion estimation.
no code implementations • 30 Jan 2019 • Simon Chadwick, Will Maddern, Paul Newman
Given the possible closing speeds between two vehicles, this necessitates the ability to accurately detect distant vehicles.
Robotics
no code implementations • 3 Jan 2019 • Horia Porav, Paul Newman
This work examines the role of reinforcement learning in reducing the severity of on-road collisions by controlling velocity and steering in situations in which contact is imminent.
no code implementations • 3 Jan 2019 • Horia Porav, Tom Bruls, Paul Newman
We present a method for improving segmentation tasks on images affected by adherent rain drops and streaks.
no code implementations • 3 Dec 2018 • Tom Bruls, Horia Porav, Lars Kunze, Paul Newman
Many tasks performed by autonomous vehicles such as road marking detection, object tracking, and path planning are simpler in bird's-eye view.
no code implementations • 11 Oct 2018 • Letizia Marchegiani, Paul Newman
This paper is about alerting acoustic event detection and sound source localisation in an urban scenario.
no code implementations • 9 Mar 2018 • Horia Porav, Will Maddern, Paul Newman
We present a method of improving visual place recognition and metric localisation under very strong appear- ance change.
no code implementations • 27 Jan 2018 • Michael Tanner, Stefan Saftescu, Alex Bewley, Paul Newman
We train a suitably deep network architecture with two 3D meshes: a high-quality laser reconstruction, and a lower quality stereo image reconstruction.
no code implementations • CVPR 2017 • Geoffrey Pascoe, Will Maddern, Michael Tanner, Pedro Pinies, Paul Newman
We propose a direct monocular SLAM algorithm based on the Normalised Information Distance (NID) metric.
no code implementations • CVPR 2018 • Paul Amayo, Pedro Pinies, Lina M. Paz, Paul Newman
Unlike greedy methods --which maximise the number of inliers-- this approach efficiently searches for a soft assignment of points to models by minimising the energy of the overall classification.
no code implementations • 13 Apr 2016 • Michael Tanner, Pedro Pinies, Lina Maria Paz, Paul Newman
The computational and memory requirements of large dense models can be prohibitive.