检索结果-内蒙古大学图书馆

Comparative Analysis of Magnetic Resonance Fingerprinting Dictionaries via Dimensionality Reduction 1

学校读者我要写书评

暂无评论

1st International Workshop on Graph Learning in Medical Imaging, GLMI 2019 held in conjunction with the 22nd International Conference on Medical Image Computing and computer-Assisted Intervention, MICCAI 2019

作者： Dzyubachyk, Oleh Koolstra, Kirsten Pezzotti, Nicola Lelieveldt, Boudewijn P. F. Webb, Andrew Börnert, Peter Division of Image Processing Department of Radiology Leiden University Medical Center Leiden Netherlands C.J. Gorter Center for High Field MRI Department of Radiology Leiden University Medical Center Leiden Netherlands Computer Graphics and Visualization Group Delft University of Technology Delft Netherlands Philips Research Eindhoven Eindhoven Netherlands Intelligent Systems Department Delft University of Technology Delft Netherlands Philips Research Hamburg Hamburg Germany

ISBN: (数字)9783030358174

ISBN: (纸本)9783030358167

Quality assessment of different Magnetic Resonance Fingerprinting (MRF) sequences and their corresponding dictionaries remains an unsolved problem. In this work we present a method in which we approach analysis of MRF dictionaries by performing dimensionality reduction and representing them as low-dimensional point sets (embeddings). Dimensionality reduction was performed using a modification of the t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm. First, we demonstrated stability of calculated embeddings that allows neglecting the stochastic nature of t-SNE. Next, we proposed and analyzed two algorithms for comparing the embeddings. Finally, we performed two simulations in which we reduced the MRF sequence/dictionary in length or size and analyzed the influence of this reduction on the resulting embedding. We believe that this research can pave the way to development of a software tool for analysis, including better understanding, optimization and comparison, of different MRF sequences. © 2019, Springer Nature Switzerland AG.

关键词： Magnetic resonance

Consistent curvature approximation on riemannian shape spaces

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Effland, Alexander Heeren, Behrend Rumpf, Martin Wirth, Benedikt Institute of Computer Graphics and Vision Graz University of Technology Graz Austria Institute for Numerical Simulation University of Bonn Bonn Germany Applied Mathematics University of Münster Münster Germany

We describe how to approximate the Riemann curvature tensor as well as sectional curvatures on possibly infinite-dimensional shape spaces that can be thought of as Riemannian manifolds. To this end, we extend the variational time discretization of geodesic calculus presented in [RW15], which just requires an approximation of the squared Riemannian distance that is typically easy to compute. First we obtain first order discrete covariant derivatives via a Schild's ladder type discretization of parallel transport. Second order discrete covariant derivatives are then computed as nested first order discrete covariant derivatives. These finally give rise to an approximation of the curvature tensor. First and second order consistency are proven for the approximations of the covariant derivative and the curvature tensor. The findings are experimentally validated on two-dimensional surfaces embedded in R3. Furthermore, as a proof of concept the method is applied to the shape space of triangular meshes, and discrete sectional curvature indicatrices are computed on low-dimensional vector *** Codes 37L65, 49M25, 53C22, 65L20, 65D18, 53C21 Copyright © 2019, The Authors. All rights reserved.

关键词： Calculations

MULTIPLE VIEW GEOMETRY IN REMOTE SENSING: AN EMPIRICAL STUDY BASED ON PLEIADES SATELLITE IMAGES

学校读者我要写书评

暂无评论

MULTIPLE VIEW GEOMETRY IN REMOTE SENSING: AN EMPIRICAL STUDY...

IEEE International Geoscience and Remote Sensing Symposium

作者： Roland Perko Mathias Schardt Livia Piermattei Stefan Auer Peter M. Roth RESEARCH Forschungsgesellschaft mbH DIGITAL Austria Department of Geodesy and Geoinformation Technical University of Vienna Austria German Aerospace Center (DLR) Remote Sensing Technology Institute Germany Institute of Computer Graphics and Vision Graz University of Technology Austria

In contrast to the fields of computer vision and photogrammetry, multiple view geometry has not been extensively exploited in the remote sensing domain so far. Therefore, an empirical study is conducted based on multi view Pleiades data that depicts a scene from multiple orbits and multiple incidence angles. First, an accuracy analysis of the 2D and 3D geo-location performance is elaborated showing that ground control points can be modelled with a root mean square residual error below 30 cm in East, North, and height. Second, digital surface models are reconstructed from all possible stereo pairs and are additionally fused in the multiple view geometry sense. It is shown that employing more data increases the accuracy of the digital surface model while reducing the amount of the non-reconstructed regions.

关键词： Sensors Geometry Image reconstruction Laser radar Remote sensing Orbits Three-dimensional displays

Reaction to directional vibrations applied on a joystick 19th

学校读者我要写书评

暂无评论

Reaction to directional vibrations applied on a joystick

19th International Conference on New Trends in Civil Aviation, NTCA 2017

作者： Zikmund, P. Macík, M. Míkovec, Z. Institute of Aerospace Engineering Faculty of Mechanical Engineering Brno University of Technology Czech Republic Department of Computer Graphics and Interaction Czech Technical University in Prague Czech Republic

ISBN: (纸本)9780815376026

Human factor is the most common reason for aircraft accident. Therefore, an improvement of pilot–aircraft interaction by vibrotactile feedback is a way of increasing the safety of flight. This study focuses on the ability of pilot to distinguish directional vibration of a control stick and react in a specific direction. To investigate the intuitive reaction of a pilot on directional vibration, human–machine experiment was carried out. For the experiment, aircraft control stick was replaced by a joystick. The task involved the reaction to the directional vibrations of segments mounted on the joystick by an intuitive reaction. A hypothesis that a human can distinguish directional vibration of the control stick and react in specific directions was confirmed with an error rate of approximately 5%. The experiment was carried out with a prescribed way of holding the joystick in hand. This fact limits results generalization. Future work aims at designing new feedback hardware and analyzing the influence of different ways of stick holding. © 2018 Taylor & Francis Group, London.

关键词： Aircraft control

Learning pose specific representations by predicting different views

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Poier, Georg Schinagl, David Bischof, Horst Institute for Computer Graphics and Vision Graz University of Technology Austria

The labeled data required to learn pose estimation for articulated objects is difficult to provide in the desired quantity, realism, density, and accuracy. To address this issue, we develop a method to learn representations, which are very specific for articulated poses, without the need for labeled training data. We exploit the observation that the object pose of a known object is predictive for the appearance in any known view. That is, given only the pose and shape parameters of a hand, the hand's appearance from any viewpoint can be approximated. To exploit this observation, we train a model that - given input from one view - estimates a latent representation, which is trained to be predictive for the appearance of the object when captured from another viewpoint. Thus, the only necessary supervision is the second view. The training process of this model reveals an implicit pose representation in the latent space. Importantly, at test time the pose representation can be inferred using only a single view. In qualitative and quantitative experiments we show that the learned representations capture detailed pose information. Moreover, when training the proposed method jointly with labeled and unlabeled data, it consistently surpasses the performance of its fully supervised counterpart, while reducing the amount of needed labeled samples by at least one order of magnitude. Copyright © 2018, The Authors. All rights reserved.

关键词： Machine learning

MURAUER: Mapping Unlabeled Real Data for Label AUstERity

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Poier, Georg Opitz, Michael Schinagl, David Bischof, Horst Institute for Computer Graphics and Vision Graz University of Technology Austria

Data labeling for learning 3D hand pose estimation models is a huge effort. Readily available, accurately labeled synthetic data has the potential to reduce the effort. However, to successfully exploit synthetic data, current state-of-the-art methods still require a large amount of labeled real data. In this work, we remove this requirement by learning to map from the features of real data to the features of synthetic data mainly using a large amount of synthetic and unlabeled real data. We exploit unlabeled data using two auxiliary objectives, which enforce that (i) the mapped representation is pose specific and (ii) at the same time, the distributions of real and synthetic data are aligned. While pose specifity is enforced by a self-supervisory signal requiring that the representation is predictive for the appearance from different views, distributions are aligned by an adversarial term. In this way, we can significantly improve the results of the baseline system, which does not use unlabeled data and outperform many recent approaches already with about 1% of the labeled real data. This presents a step towards faster deployment of learning based hand pose estimation, making it accessible for a larger range of applications. Copyright © 2018, The Authors. All rights reserved.

关键词： Machine learning

Vision based vehicle relocalization in 3D line-feature map using Perspective-n-Line with a known vertical direction

学校读者我要写书评

暂无评论

Vision based vehicle relocalization in 3D line-feature map u...

International Conference on Intelligent Transportation

作者： Louis Lecrosnier Rémi Boutteau Pascal Vasseur Xavier Savatier Friedrich Fraundorfer Normandie Univ UNIROUEN ESIGELEC IRSEEM Rouen France Normandie Univ UNIROUEN UNIHAVRE INSA Rouen LITIS Rouen France Institute for Computer Graphics and Vision Graz University of Technology Graz Austria

ISBN: (纸本)9781538670255

Common approaches for vehicle localization pro-pose to match LiDAR data or 2D features from cameras to a prior 3D LiDAR map. Yet, these methods require both heavy computational power often provided by GPU, and a first rough localization estimate via GNSS to be performed online. Moreover, storing and accessing 3D dense LiDAR maps can be challenging in case of city-wide *** this paper, we address the problem of camera global relocalization in a prior 3D line-feature map from a single image, in a GNSS denied context and with no prior pose estimation. We propose a dual contribution.(1) We introduce a novel pose estimation method from lines, (i.e. Perspective-n-Line or PnL), with a known vertical direction. Our method benefits a Gauss-Newton optimization scheme to compensate the sensor-induced vertical direction errors, and refine the overall pose. Our algorithm requires at least 3 lines to output a pose (P3L) and requires no reformulation to operate with a higher number of lines.(2) We propose a RANSAC (RANdom SAmple Consensus) 2D-3D line matching and outliers removal algorithm requiring solely one 2D-3D line pair to operate, i.e. RANSAC1. Our method reduces the number of iteration required to match features and can be easily modified to exhaustively test all feature *** evaluate the robustness of our algorithms with a synthetic data, and on a challenging sub-sequence of the KITTI dataset.

关键词： Three-dimensional displays Cameras Laser radar Pose estimation Two dimensional displays Global navigation satellite system Feature extraction

The quest for the golden activation function

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Basirat, Mina Roth, Peter M. Graz University of Technology Institute of Computer Graphics and Vision Austria

Deep Neural Networks have been shown to be beneficial for a variety of tasks, in particular allowing for end-to-end learning and reducing the requirement for manual design decisions. However, still many parameters have to be chosen in advance, also raising the need to optimize them. One important, but often ignored system parameter is the selection of a proper activation function. Thus, in this paper we target to demonstrate the importance of activation functions in general and show that for different tasks different activation functions might be meaningful. To avoid the manual design or selection of activation functions, we build on the idea of genetic algorithms to learn the best activation function for a given task. In addition, we introduce two new activation functions, ELiSH and HardELiSH, which can easily be incorporated in our framework. In this way, we demonstrate for three different image classification benchmarks that different activation functions are learned, also showing improved results compared to typically used baselines. Copyright © 2018, The Authors. All rights reserved.

关键词： Deep neural networks

Automation of processor verification using recurrent neural networks

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Fajcik, Martin Smrz, Pavel Zachariasova, Marcela Department of Computer Graphics and Multimedia Brno University of Technology Department of Computer Systems Brno University of Technology

When considering simulation-based verification of processors, the current trend is to generate stimuli using pseudorandom generators (PRGs), apply them to the processor inputs and monitor the achieved coverage of its functionality in order to determine verification completeness. Stimuli can have different forms, for example, they can be represented by bit vectors applied to the input ports of the processor or by programs that are loaded directly into the program memory. In this paper, we propose a new technique dynamically altering constraints for PRG via recurrent neural network, which receives a coverage feedback from the simulation of design under verification. For the demonstration purposes we used processors provided by Codasip as their coverage state space is reasonably big and differs for various kinds of processors. Nevertheless, techniques presented in this paper are widely applicable. The results of experiments show that not only the coverage closure is achieved much sooner, but we are able to isolate a small set of stimuli with high coverage that can be used for running regression tests. Copyright © 2018, The Authors. All rights reserved.

关键词： Combinatorial optimization