检索结果-内蒙古大学图书馆

Same Same but Different: Augmentation of Tiny Industrial Datasets using Generative Adversarial Networks

学校读者我要写书评

暂无评论

Same Same but Different: Augmentation of Tiny Industrial Dat...

Swiss Conference on Data Science (SDS)

作者： Evangelos Ntavelis Iason Kastanis Luc Van Gool Radu Timofte CVL ETH Zurich Robotics & ML CSEM SA Zurich Switzerland Robotics & ML CSEM SA Alpnach Switzerland CVL ETH Zurich PSI ESAT KU Leuven Zurich Switzerland Computer Vision Lab D-ITET ETH Zurich Zurich Switzerland

ISBN: (数字)9781728171777

ISBN: (纸本)9781728171784

The evolution of generative adversarial networks has permitted the generation of realistic fake images which, in some cases, are indistinguishable from the real ones. Many recent works in image generation focus on learning internal image statistics via training only on a single natural image. While natural images exhibit a variability in their attributes, industrial images are often acquired in a controlled environment following a specific structure. In this work we utilize the cutting-edge results of single image generation on the structured case of industrial images. Deep Learning plays an important role in Industry 4.0 manufacturing lines and multiple ML-based image processing products are currently on the market. To be able to tackle a variety of problems where image acquisition is costly and time-consuming data generation is a promising approach. The proposed method only requires a handful of images for training, making it an ideal candidate for industrial application where data is scarce and confidential. It provides the foundation for a variety of use cases in the field of industrial Deep Learning.

关键词： Generators Image resolution Training Gallium nitride Image generation Generative adversarial networks Robots

SkyCam: A dataset of sky images and their irradiance values

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Ntavelis, Evangelos Remund, Jan Schmid, Philipp Robotics & Deep Learning CSEM SA Computer Vision Lab ETH Zurich Alpnach Switzerland Energy & Climate Meteotest Bern Switzerland Industry 4.0 & Machine Learning CSEM SA Alpnach Switzerland

Recent advances in computer vision and Deep Learning have enabled astonishing results in a variety of fields and applications. Motivated by this success, the SkyCam Dataset aims to enable image-based Deep Learning solutions for short-term, precise prediction of solar radiation on a local level. For the span of a year, three different cameras in three topographically different locations in Switzerland are acquiring images of the sky every 10 seconds. Thirteen high resolution images with different exposure times are captured and used to create an additional HDR image. The images are paired with highly precise irradiance values gathered from a high-accuracy pyranometer. © 2021, CC BY-NC-SA.

关键词： Deep learning

A Real-Time Image Stitching Framework for Fetoscopic Field-of-View Expansion

学校读者我要写书评

暂无评论

A Real-Time Image Stitching Framework for Fetoscopic Field-o...

IEEE International Symposium on Biomedical Imaging

作者： Rowan Honeywell Radian Gondokaryono Rory Windrim Lueder A. Kahrs Medical Computer Vision and Robotics Lab University of Toronto Toronto Canada Ontario Fetal Center at Mount Sinai Hospital Toronto Canada Division of Maternal-Fetal Medicine University of Toronto Toronto Canada

ISBN: (数字)9798350313338

ISBN: (纸本)9798350313345

Twin-to-twin transfusion syndrome is a serious prenatal condition whose treatment employs the use of a fetoscope and optical laser for targeted vessel ablation. During surgery, challenges such as a limited field-of-view and inconsistent lighting can be addressed by field-of-view expansion. As foundation, we draw upon a state-of-the-art method that utilizes a U-Net segmentation model to mitigate visible in utero inconsistencies. We adapt this method for real-time performance by implementing a feature-based stitching algorithm. The proposed framework includes a loop closure and map correction procedure to help reduce the accumulated drifting error. During evaluation on simulated fetoscopic movements, the chosen feature extractor, SIFT, permits an average homography estimation time of 0.07s and a failure rate of 4%. Homography parameters such as scaling, translation, and rotation are isolated to identify obstacles caused by certain camera movements. The framework is evaluated using two phantom placentas in different environmental conditions and submerged underwater. Qualitative and quantitative results are presented as well as an analysis of the map correction procedure. These evaluations help recognize limitations of the framework that can be investigated and addressed in future research.

关键词： Image segmentation Surgery Phantoms Lighting Feature extraction Optical imaging Laser modes

Policy-Guided Lazy Search with Feedback for Task and Motion Planning

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Khodeir, Mohamed Sonwane, Atharv Hari, Ruthrash Shkurti, Florian Robot Vision and Learning Lab University of Toronto Robotics Institute Canada Dept. of Computer Science University of BITS Pilani India Dept. of Mathematical and Computational Sciences University of Toronto Canada

PDDLStream solvers have recently emerged as viable solutions for Task and Motion Planning (TAMP) problems, extending PDDL to problems with continuous action spaces. Prior work has shown how PDDLStream problems can be reduced to a sequence of PDDL planning problems, which can then be solved using off-the-shelf planners. However, this approach can suffer from long runtimes. In this paper we propose LAZY, a solver for PDDLStream problems that maintains a single integrated search over action skeletons, which gets progressively more geometrically informed, as samples of possible motions are lazily drawn during motion planning. We explore how learned models of goal-directed policies and current motion sampling data can be incorporated in LAZY to adaptively guide the task planner. We show that this leads to significant speed-ups in the search for a feasible solution evaluated over unseen test environments of varying numbers of objects, goals, and initial conditions. We evaluate our TAMP approach by comparing to existing solvers for PDDLStream problems on a range of simulated 7DoF rearrangement/manipulation problems. Code can be found at https://***/ learning-based-tamp. © 2022, CC BY.

关键词： Motion planning

The area bisectors of a polygon and force equilibria in programmable vector fields 97

学校读者我要写书评

暂无评论

The area bisectors of a polygon and force equilibria in prog...

Proceedings of the thirteenth annual symposium on Computational geometry

作者： Karl-Friedrich Böhringer Bruce Randall Donald Dan Halperin ALPHA Lab Dept. of Ind. Eng. and Op. Research Univ. of CA Berkeley and Robotics & Vision Lab Dept. of Comp. Sc. Cornell Univ. Robotics & Vision Laboratory Department of Computer Science Cornell University Dept. of Comp. Sc. Tel Aviv Univ. Tel Aviv 69978 Israel

Towards efficient human-robot cooperation for socially-aware robot navigation in human-populated environments: the SNAPE framework

学校读者我要写书评

暂无评论

Towards efficient human-robot cooperation for socially-aware...

IEEE International Conference on robotics and Automation (ICRA)

作者： A. Vega-Magro R. Gondkar L.J. Manso P. Núñez Robotics and Artificial Vision Lab. RoboLab Group University of Extremadura Spain Pune Institute of Computer Technology Pune University India College of Engineering and Physical Sciences Aston University Birmingham UK

It is widely accepted that in the future, robots will cooperate with humans in everyday tasks. Robots interacting with humans will require social awareness when performing their tasks which will require navigation. While navigating, robots should aim to avoid distressing people in order to maximize their chance of social acceptance. For instance, avoiding getting too close to people or disrupting interactions. Most research approaches these problems by planning socially accepted paths, however, in everyday situations, there are many examples where a simple path planner cannot solve all of the predicted robots’ navigation problems. For instance, requesting permission to interrupt a conversation if an alternative path cannot be determined requires deliberative skills. This article presents the Social Navigation framework for Autonomous robots in Populated Environments (SNAPE), where different software agents are integrated within a robotics cognitive architecture. SNAPE addresses action planning aimed at social-awareness navigation in realistic situations: it plans socially accepted paths and conversations to negotiate its trajectory to reach targets. In this article, the framework is evaluated in different use-cases where the robot, during its navigation, has to interact with different people in order to reach its goal. The results show that participants report that the robot’s behavior was realistic and human-like.

关键词： Navigation Conferences computer architecture Software agents Planning Trajectory Task analysis

3D Large-Scale Seabed Reconstruction for UUV Simultaneous Localization and Mapping

学校读者我要写书评

暂无评论

IFAC Proceedings Volumes 2008年第1期41卷 19-24页

作者： Yvan R. Petillot Joaquim Salvi Elisabet Batlle Ocean Systems Lab Heriot-Watt University EH14-4AS Edinburgh Scotland (Tel: +44 (0)131 451 8277) Computer Vision and Robotics University of Girona E-17071 Girona Spain (Tel: +34 972 41 8483)

This paper proposes a new technique to reconstruct large 3D scenes from a sequence of video images by combining Bayesian filtering and state-of-art 3D computer vision. The approach performs the alignment of a sequence of 3D partial reconstructions of the seafloor thanks to the re-observations of passive landmarks by means of a linear Kalman filter-based SLAM approach. Landmarks are detected on the images and characterized considering 2D and 3D features. Landmarks are re-observed while the robot is navigating and data association becomes easier but robust. Preliminary results are performed in virtual scenarios but processing real images synthetized from underwater textures.

关键词： Visual SLAM Surface Registration Stereo vision Underwater Imaging computer vision Simultaneous Localization and Mapping

Convolutional Cross-View Pose Estimation

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Xia, Zimin Booij, Olaf Kooij, Julian F.P. The Intelligent Vehicles Group Department of Cognitive Robotics Delft University of Technology Leeghwaterstraat Delft2628 CN Netherlands The Computer Vision Lab Delft University of Technology Mekelweg 4 Delft2628 CD Netherlands

We propose a novel end-to-end method for cross-view pose estimation. Given a ground-level query image and an aerial image that covers the query’s local neighborhood, the 3 Degrees-of-Freedom camera pose of the query is estimated by matching its image descriptor to descriptors of local regions within the aerial image. The orientation-aware descriptors are obtained by using a translationally equivariant convolutional ground image encoder and contrastive learning. The Localization Decoder produces a dense probability distribution in a coarse-to-fine manner with a novel Localization Matching Upsampling module. A smaller Orientation Decoder produces a vector field to condition the orientation estimate on the localization. Our method is validated on the VIGOR and KITTI datasets, where it surpasses the state-of-the-art baseline by 72% and 36% in median localization error for comparable orientation estimation accuracy. The predicted probability distribution can represent localization ambiguity, and enables rejecting possible erroneous predictions. Without re-training, the model can infer on ground images with different field of views and utilize orientation priors if available. On the Oxford RobotCar dataset, our method can reliably estimate the ego-vehicle’s pose over time, achieving a median localization error under 1 meter and a median orientation error of around 1 degree at 14 FPS. Copyright © 2023, The Authors. All rights reserved.

关键词： Cameras

Learning to avoid poor images: Towards task-aware C-arm cone-beam CT trajectories

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Zaech, Jan-Nico Gao, Cong Bier, Bastian Taylor, Russell Maier, Andreas Navab, Nassir Unberath, Mathias Laboratory for Computational Sensing and Robotics Johns Hopkins University Pattern Recognition Lab Friedrich-Alexander-Universität Erlangen-Nürnberg Computer Vision Laboratory Eidgenössische Technische Hochschule Zürich

Metal artifacts in computed tomography (CT) arise from a mismatch between physics of image formation and idealized assumptions during tomographic reconstruction. These artifacts are particularly strong around metal implants, inhibiting widespread adoption of 3D cone-beam CT (CBCT) despite clear opportunity for intra-operative verification of implant positioning, e. g. in spinal fusion surgery. On synthetic and real data, we demonstrate that much of the artifact can be avoided by acquiring better data for reconstruction in a task-aware and patient-specific manner, and describe the first step towards the envisioned task-aware CBCT protocol. The traditional short-scan CBCT trajectory is planar, with little room for scene-specific adjustment. We extend this trajectory by autonomously adjusting out-of-plane angulation. This enables C-arm source trajectories that are scene-specific in that they avoid acquiring"poor images", characterized by beam hardening, photon starvation, and noise. The recommendation of ideal out-of-plane angulation is performed on-the-fly using a deep convolutional neural network that regresses a detectability-rank derived from imaging physics. Copyright © 2019, The Authors. All rights reserved.

关键词： Trajectories