arXiv subjects

Wen Yang

Publications and source records attributed to Wen Yang.

At least 145 records · Page 8Linked to original sources

arXiv2022

Targetless Extrinsic Calibration of Stereo Cameras, Thermal Cameras, and Laser Sensors in the Wild

The fusion of multi-modal sensors has become increasingly popular in autonomous driving and intelligent robots since it can provide richer information than any single sensor, enhance reliability in complex environments. Multi-sensor extrinsic calibration is one of the key factors of sensor fusion. However, such calibration is difficult due to the variety of sensor modalities and the requirement of calibration targets and human labor. In this paper, we demonstrate a new targetless cross-modal calibration framework by focusing on the extrinsic transformations among stereo cameras, thermal cameras, and laser sensors. Specifically, the calibration between stereo and laser is conducted in 3D space by minimizing the registration error, while the thermal extrinsic to the other two sensors is estimated by optimizing the alignment of the edge features. Our method requires no dedicated targets and performs the multi-sensor calibration in a single shot without human interaction. Experimental results show that the calibration framework is accurate and applicable in general scenes.

cs.RO↗

arXiv2022

Graph-Based Similarity of Neural Network Representations

Understanding the black-box representations in Deep Neural Networks (DNN) is an essential problem in deep learning. In this work, we propose Graph-Based Similarity (GBS) to measure the similarity of layer features. Contrary to previous works that compute the similarity directly on the feature maps, GBS measures the correlation based on the graph constructed with hidden layer outputs. By treating each input sample as a node and the corresponding layer output similarity as edges, we construct the graph of DNN representations for each layer. The similarity between graphs of layers identifies the correspondences between representations of models trained in different datasets and initializations. We demonstrate and prove the invariance property of GBS, including invariance to orthogonal transformation and invariance to isotropic scaling, and compare GBS with CKA. GBS shows state-of-the-art performance in reflecting the similarity and provides insights on explaining the adversarial sample behavior on the hidden layer space.

cs.LG↗

arXiv2022

Learning to Extract Building Footprints from Off-Nadir Aerial Images

Extracting building footprints from aerial images is essential for precise urban mapping with photogrammetric computer vision technologies. Existing approaches mainly assume that the roof and footprint of a building are well overlapped, which may not hold in off-nadir aerial images as there is often a big offset between them. In this paper, we propose an offset vector learning scheme, which turns the building footprint extraction problem in off-nadir images into an instance-level joint prediction problem of the building roof and its corresponding "roof to footprint" offset vector. Thus the footprint can be estimated by translating the predicted roof mask according to the predicted offset vector. We further propose a simple but effective feature-level offset augmentation module, which can significantly refine the offset vector prediction by introducing little extra cost. Moreover, a new dataset, Buildings in Off-Nadir Aerial Images (BONAI), is created and released in this paper. It contains 268,958 building instances across 3,300 aerial images with fully annotated instance-level roof, footprint, and corresponding offset vector for each building. Experiments on the BONAI dataset demonstrate that our method achieves the state-of-the-art, outperforming other competitors by 3.37 to 7.39 points in F1-score. The codes, datasets, and trained models are available at https://github.com/jwwangchn/BONAI.git.

cs.CV↗

arXiv2022

A Survey of Uncertainty in Deep Neural Networks

Due to their increasing spread, confidence in neural network predictions became more and more important. However, basic neural networks do not deliver certainty estimates or suffer from over or under confidence. Many researchers have been working on understanding and quantifying uncertainty in a neural network's prediction. As a result, different types and sources of uncertainty have been identified and a variety of approaches to measure and quantify uncertainty in neural networks have been proposed. This work gives a comprehensive overview of uncertainty estimation in neural networks, reviews recent advances in the field, highlights current challenges, and identifies potential research opportunities. It is intended to give anyone interested in uncertainty estimation in neural networks a broad overview and introduction, without presupposing prior knowledge in this field. A comprehensive introduction to the most crucial sources of uncertainty is given and their separation into reducible model uncertainty and not reducible data uncertainty is presented. The modeling of these uncertainties based on deterministic neural networks, Bayesian neural networks, ensemble of neural networks, and test-time data augmentation approaches is introduced and different branches of these fields as well as the latest developments are discussed. For a practical application, we discuss different measures of uncertainty, approaches for the calibration of neural networks and give an overview of existing baselines and implementations. Different examples from the wide spectrum of challenges in different fields give an idea of the needs and challenges regarding uncertainties in practical applications. Additionally, the practical limitations of current methods for mission- and safety-critical real world applications are discussed and an outlook on the next steps towards a broader usage of such methods is given.

cs.LG↗

arXiv2022

First-principles study on the solute-induced low diffusion and self-trapping of helium in fcc iron

The addition of alloying elements plays an essential role in helium (He) behaviours produced by transmutation in metal alloys. Effects of solutes (Ni, Cr, Ti, P, Si, C) on the behaviours of He and He-He pair in face-centred cube (fcc) iron have been investigated using first-principles calculations based on density functional theory (DFT). For the interactions of solutes and He, we found that Ti, P, Si, and C attracts He is more potent than Ni and Cr in fcc iron. We have determined the most stable configuration for the He-He pair, which is the Hesub-Hetetra pair with a binding energy of 1.60 eV. In considering the effect of solutes on the stability of the He-He pair, we have proposed a unique definition of binding energy. By applying the definition, we suggest that Ti and P could weaken He self-trapping, and Cr and C are beneficial for He self-trapping, while Ni is similar to the matrix Fe itself. For the diffusion of He, which is the necessary process of forming the He bubble, we determined that the most stable interstitial He is in a tetrahedral site and could migrate with the energy barrier of 0.16 eV in pure fcc iron. We further found that Ti and Si can increase the barrier to 0.18 and 0.20 eV; on the contrary, Cr and P decrease the barrier to 0.10 and 0.06 eV, respectively. Summarizing the calculations, we conclude that Ti decreases while Cr increases the diffusion and self-trapping of He in fcc iron.

cond-mat.mtrl-sci↗

arXiv2021

Generic properties of free boundary problems in plasma physics

We are concerned with the global bifurcation analysis of positive solutions to free boundary problems arising in plasma physics. We show that in general, in the sense of domain variations, the following alternative holds: either the shape of the branch of solutions resembles the monotone one of the model case of the two-dimensional disk, or it is a continuous simple curve without bifurcation points which ends up at a point where the boundary density vanishes. On the other hand, we deduce a general criterion ensuring the existence of a free boundary in the interior of the domain. Application to a classic nonlinear eigenvalue problem is also discussed.

math.AP↗

arXiv2021

Object Detection in Aerial Images: A Large-Scale Benchmark and Challenges

In the past decade, object detection has achieved significant progress in natural images but not in aerial images, due to the massive variations in the scale and orientation of objects caused by the bird's-eye view of aerial images. More importantly, the lack of large-scale benchmarks has become a major obstacle to the development of object detection in aerial images (ODAI). In this paper,we present a large-scale Dataset of Object deTection in Aerial images (DOTA) and comprehensive baselines for ODAI. The proposed DOTA dataset contains 1,793,658 object instances of 18 categories of oriented-bounding-box annotations collected from 11,268 aerial images. Based on this large-scale and well-annotated dataset, we build baselines covering 10 state-of-the-art algorithms with over 70 configurations, where the speed and accuracy performances of each model have been evaluated. Furthermore, we provide a code library for ODAI and build a website for evaluating different algorithms. Previous challenges run on DOTA have attracted more than 1300 teams worldwide. We believe that the expanded large-scale DOTA dataset, the extensive baselines, the code library and the challenges can facilitate the designs of robust algorithms and reproducible research on the problem of object detection in aerial images.

cs.CV↗

arXiv2021

An F-modulated stability framework for multistep methods

We introduce a new $\mathbf F$-modulated energy stability framework for general linear multistep methods. We showcase the theory for the two dimensional molecular beam epitaxy model with no slope selection which is a prototypical gradient flow with Lipschitz-bounded nonlinearity. We employ a class of representative BDF$k$, $2\le k \le 5$ discretization schemes with explicit $k^{\mathrm{th}}$-order extrapolation of the nonlinear term. We prove the uniform-in-time boundedness of high Sobolev norms of the numerical solution. The upper bound is unconditional, i.e. regardless of the size of the time step. We develop a new algebraic theory and calibrate nearly optimal and \emph{explicit} maximal time step constraints which guarantee monotonic $\mathbf F$-modulated energy dissipation.

math.NA↗

arXiv2021

From KP-I lump solution to travelling waves of Gross-Pitaevskii equation

Let $q(x,y)$ be an nondegenerate lump solution to KP-I (Kadomtsev-Petviashvili-I) equation $$\partial_x^4q-2\sqrt{2}\partial_x^2q-3\sqrt{2}\partial_x((\partial_xq) ^2)-2\partial_y^2q=0. $$ We prove the existence of a traveling wave solution $ u_{\e} (x-ct, y)$ to GP (Gross-Pitaevskii) equation $$ i\partial_{t}Ψ+ΔΨ+(1-|Ψ|^{2})Ψ=0,\ \ \ \mbox{in} \ {\mathbb R}^2 $$ in the transonic limit $$ c=\sqrt{2}-ε^2 $$ with $$ u_ε=1 + i εq(x,y) + {\mathcal O} (ε^2). $$ This proves the existence of finite energy solutions in the so-called Jones-Roberts program in the transonic range $ c \in (\sqrt{2}-ε^2, \sqrt{2})$. The main ingredients in our proof are detailed point-wise estimates of the Green function associated to a family of fourth order hypoelliptic operators $$\partial_x^4-(2\sqrt{2}-\e^2)\partial_x^2-2\partial_y^2+\e^2\partial_x^2\partial_y^2+\e^4\partial_y^4.$$

math.AP↗

arXiv2021

The BDF3/EP3 scheme for MBE with no slope selection is stable

We consider the classical molecular beam epitaxy (MBE) model with logarithmic type potential known as no-slope-selection. We employ a third order backward differentiation (BDF3) in time with implicit treatment of the surface diffusion term. The nonlinear term is approximated by a third order explicit extrapolation (EP3) formula. We exhibit mild time step constraints under which the modified energy dissipation law holds. We break the second Dahlquist barrier and develop a new theoretical framework to prove unconditional uniform energy boundedness with no size restrictions on the time step. This is the first unconditional result for third order BDF methods applied to the MBE models without introducing any stabilization terms or fictitious variables. A novel theoretical framework is also established for the error analysis of high order methods.

math.NA↗

arXiv2021

Motion Deblurring with Real Events

In this paper, we propose an end-to-end learning framework for event-based motion deblurring in a self-supervised manner, where real-world events are exploited to alleviate the performance degradation caused by data inconsistency. To achieve this end, optical flows are predicted from events, with which the blurry consistency and photometric consistency are exploited to enable self-supervision on the deblurring network with real-world data. Furthermore, a piece-wise linear motion model is proposed to take into account motion non-linearities and thus leads to an accurate model for the physical formation of motion blurs in the real-world scenario. Extensive evaluation on both synthetic and real motion blur datasets demonstrates that the proposed algorithm bridges the gap between simulated and real-world motion blurs and shows remarkable performance for event-based motion deblurring in real-world scenarios.

cs.CV↗

arXiv2021

Toward Efficient and Robust Multiple Camera Visual-inertial Odometry

Efficiency and robustness are the essential criteria for the visual-inertial odometry (VIO) system. To process massive visual data, the high cost on CPU resources and computation latency limits VIO's possibility in integration with other applications. Recently, the powerful embedded GPUs have great potentials to improve the front-end image processing capability. Meanwhile, multi-camera systems can increase the visual constraints for back-end optimization. Inspired by these insights, we incorporate the GPU-enhanced algorithms in the field of VIO and thus propose a new front-end with NVIDIA Vision Programming Interface (VPI). This new front-end then enables multi-camera VIO feature association and provides more stable back-end pose optimization. Experiments with our new front-end on monocular datasets show the CPU resource occupation rate and computational latency are reduced by 40.4% and 50.6% without losing accuracy compared with the original VIO. The multi-camera system shows a higher VIO initialization success rate and better robustness overall state estimation.

cs.RO↗

arXiv2021

Implementation of weak measurement amplification with the weak coherent light and optomechanical system

Most studies for postselected weak measurement in optomechanical system focus on using a single photon as a measured system. However, we find that using weak coherent light instead of a single photon can also amplify the mirror's position displacement of one photon. In the WVA regime (Weak Value Amplification regime), the weak value of one photon can lie outside the eigenvalue spectrum, proportional to the difference of the mirror's position displacement between the successful and failed postselection and its successful postselection probability is dependent of the mean photon number of the coherent light and improved by adjusting it accordingly. Outside the WVA regime, the amplification limit can reach the level of the vacuum fluctuations, and when the mean photon number and the optomechanical coupling parameter are selected appropriately, its successful postselection probability becomes higher, which is beneficial to observe the maximum amplification value under the current experimental conditions. This result breaks the constraint that it is difficult to detect outside the WVA regime. This opens up a new regime for the study of a single photon nonlinearity in optomechanical system.

quant-ph↗

arXiv2021

MD Simulation of Hundred-Billion-Metal-Atom Cascade Collision on Sunway Taihulight

Radiation damage to the steel material of reactor pressure vessels is a major threat to the nuclear reactor safety. It is caused by the metal atom cascade collision, initialized when the atoms are struck by a high-energy neutron. The paper presents MISA-MD, a new implementation of molecular dynamics, to simulate such cascade collision with EAM potential. MISA-MD realizes (1) a hash-based data structure to efficiently store an atom and find its neighbors, and (2) several acceleration and optimization strategies based on SW26010 processor of Sunway Taihulight supercomputer, including an efficient potential table storage and interpolation method, a coloring method to avoid write conflicts, and double-buffer and data reuse strategies. The experimental results demonstrated that MISA-MD has good accuracy and scalability, and obtains a parallel efficiency of over 79% in an 655-billion-atom system. Compared with a state-of-the-art MD program LAMMPS, MISA-MD requires less memory usage and achieves better computational performance.

cs.DC↗

arXiv2021

On a sinc-type MBE model

We introduce a new sinc-type molecular beam epitaxy model which is derived from a cosine-type energy functional. The landscape of the new functional is remarkably similar to the classical MBE model with double well potential but has the additional advantage that all its derivatives are uniformly bounded. We consider first order IMEX and second order BDF2 discretization schemes. For both cases we quantify explicit time step constraints for the energy dissipation which is in good accord with the practical numerical simulations. Furthermore we introduce a new theoretical framework and prove unconditional uniform energy boundedness with no size restrictions on the time step. This is the first unconditional (i.e. independent of the time step size) result for semi-implicit methods applied to the phase field models without introducing any artificial stabilization terms or fictitious variables.

math.NA↗

arXiv2021

Sharp convergence to steady states of Allen-Cahn

In our recent work we found a surprising breakdown of symmetry conservation: using standard numerical discretization with very high precision the computed numerical solutions corresponding to very nice initial data may converge to completely incorrect steady states due to the gradual accumulation of machine round-off error. We solved this issue by introducing a new Fourier filter technique for solutions with certain band gap properties. To further investigate the attracting basin of steady states we classify in this work all possible bounded nontrivial steady states for the Allen-Cahn equation. We characterize sharp dependence of nontrivial steady states on the diffusion coefficient and prove strict monotonicity of the associated energy. In particular, we establish a certain self-replicating property amongst the hierarchy of steady states and give a full classification of their energies and profiles. We develop a new modulation theory and prove sharp convergence to the steady state with explicit rates and profiles.

math.AP↗

arXiv2021

On a parabolic sine-Gordon model

We consider a parabolic sine-Gordon model with periodic boundary conditions. We prove a fundamental maximum principle which gives a priori uniform control of the solution. In the one-dimensional case we classify all bounded steady states and exhibit some explicit solutions. For the numerical discretization we employ first order IMEX, and second order BDF2 discretization without any additional stabilization term. We rigorously prove the energy stability of the numerical schemes under nearly sharp and quite mild time step constraints. We demonstrate the striking similarity of the parabolic sine-Gordon model with the standard Allen-Cahn equations with double well potentials.

math.AP↗

arXiv2021

On symmetry breaking of Allen-Cahn

We consider numerical solutions for the Allen-Cahn equation with standard double well potential and periodic boundary conditions. Surprisingly it is found that using standard numerical discretizations with high precision computational solutions may converge to completely incorrect steady states. This happens for very smooth initial data and state-of-the-art algorithms. We analyze this phenomenon and showcase the resolution of this problem by a new symmetry-preserving filter technique. We develop a new theoretical framework and rigorously prove the convergence to steady states for the filtered solutions.

math.AP↗