Collaborative Augmented Reality on Smartphones via Life-long City-scale Maps

Platinsky, Lukas; Szabados, Michal; Hlasek, Filip; Hemsley, Ross; Pero, Luca Del; Pancik, Andrej; Baum, Bryan A.; Grimmett, Hugo; Ondrúška, Peter

doi:10.1109/ismar50242.2020.00081

Cited by 22 publications

(20 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also collected videos in Paris and Shanghai areas using dashcam cameras. We recovered ground truth poses from videos thanks to a large scale 3D reconstruction pipeline based on colmap SfM software [25], similar to [19]. We provide examples from these datasets in figure 6 and we briefly introduce them:…”

Section: B Results On Large-scale Datasetsmentioning

confidence: 99%

CoordiNet: uncertainty-aware pose regressor for reliable vehicle localization

Moreau¹,

Piasco²,

Tsishkou³

et al. 2021

Preprint

View full text Add to dashboard Cite

In this paper, we investigate visual-based camera localization with neural networks for robotics and autonomous vehicles applications. Our solution is a CNN-based algorithm which predicts camera pose (3D translation and 3D rotation) directly from a single image. It also provides an uncertainty estimate of the pose. Pose and uncertainty are learned together with a single loss function. Furthermore, we propose a new fully convolutional architecture, named CoordiNet, designed to embed some of the scene geometry.Our framework outperforms comparable methods on the largest available benchmark, the Oxford RobotCar dataset, with an average error of 8 meters where previous best was 19 meters. We have also investigated the performance of our method on large scenes for real time (18 fps) vehicle localization. In this setup, structure-based methods require a large database, and we show that our proposal is a reliable alternative, achieving 29cm median error in a 1.9km loop in a busy urban area.

show abstract

Section: B Results On Large-scale Datasetsmentioning

confidence: 99%

CoordiNet: uncertainty-aware pose regressor for reliable vehicle localization

Moreau¹,

Piasco²,

Tsishkou³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…This data provides us with a full 360 • field of view and range well over 100 meters. For commodity vision data, we use a visual mapping and localisation system [34], and a stereo detector dynamic object perception system [26]. We build upon DSGN [28] for the 3D object detection task.…”

Section: E Training Data: Expensive Hd Vs Commodity Visionmentioning

confidence: 99%

Quantity over Quality: Training an AV Motion Planner with Large Scale Commodity Vision Data

Platinsky¹,

Naseer²,

Chen³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

With the Autonomous Vehicle (AV) industry shifting towards Autonomy 2.0 [1], the performance of self-driving systems starts to rely heavily on large quantities of expert driving demonstrations. However, collecting this demonstration data typically involves expensive HD sensor suites (LiDAR + RADAR + cameras), which quickly becomes financially infeasible at the scales required. This motivates the use of commodity vision sensors for data collection, which are an order of magnitude cheaper than the HD sensor suites, but offer lower fidelity. If it were possible to leverage these for training an AV motion planner, observing the 'long tail' of driving events would become a financially viable strategy.As our main contribution we show it is possible to train a high-performance motion planner using commodity vision data which outperforms planners trained on HD-sensor data for a fraction of the cost. We do this by comparing the autonomy system performance when training on these two different sensor configurations, and showing that we can compensate for the lower sensor fidelity by means of increased quantity: a planner trained on 100h of commodity vision data outperforms one with 25h of expensive HD data (see Fig. 1). We also share the technical challenges we had to tackle to make this work. To the best of our knowledge, we are the first to demonstrate that this is possible using real-world data.

show abstract

“…A state of the art that highlights the intersections between AR and computer-supported collaborative work and related future research scenarios can be found in [138]. Some of these applications are dedicated to the navigation of architecture, interacting with other users who experience the same simulation [139]. In the AEC realm, collaboration on building monitoring and verification is increasingly essential, as highlighted in [140].…”

Section: Collaborative Interfacesmentioning

confidence: 99%

AR in the Architecture Domain: State of the Art

Russo

2021

Applied Sciences

View full text Add to dashboard Cite

Augmented reality (AR) allows the real and digital worlds to converge and overlap in a new way of observation and understanding. The architectural field can significantly benefit from AR applications, due to their systemic complexity in terms of knowledge and process management. Global interest and many research challenges are focused on this field, thanks to the conjunction of technological and algorithmic developments from one side, and the massive digitization of built data. A significant quantity of research in the AEC and educational fields describes this state of the art. Moreover, it is a very fragmented domain, in which specific advances or case studies are often described without considering the complexity of the whole development process. The article illustrates the entire AR pipeline development in architecture, from the conceptual phase to its application, highlighting each step’s specific aspects. This storytelling aims to provide a general overview to a non-expert, deepening the topic and stimulating a democratization process. The aware and extended use of AR in multiple areas of application can lead a new way forward for environmental understanding, bridging the gap between real and virtual space in an innovative perception of architecture.

show abstract

Collaborative Augmented Reality on Smartphones via Life-long City-scale Maps

Cited by 22 publications

References 29 publications

CoordiNet: uncertainty-aware pose regressor for reliable vehicle localization

CoordiNet: uncertainty-aware pose regressor for reliable vehicle localization

Quantity over Quality: Training an AV Motion Planner with Large Scale Commodity Vision Data

AR in the Architecture Domain: State of the Art

Contact Info

Product

Resources

About