Searcharxiv⌕ Search

arXiv subjects

Zhiqiang Li

Publications and source records attributed to Zhiqiang Li.

At least 73 records · Page 4Linked to original sources
arXiv2022

TransLoc3D : Point Cloud based Large-scale Place Recognition using Adaptive Receptive Fields

Place recognition plays an essential role in the field of autonomous driving and robot navigation. Point cloud based methods mainly focus on extracting global descriptors from local features of point clouds. Despite having achieved promising results, existing solutions neglect the following aspects, which may cause performance degradation: (1) huge size difference between objects in outdoor scenes; (2) moving objects that are unrelated to place recognition; (3) long-range contextual information. We illustrate that the above aspects bring challenges to extracting discriminative global descriptors. To mitigate these problems, we propose a novel method named TransLoc3D, utilizing adaptive receptive fields with a point-wise reweighting scheme to handle objects of different sizes while suppressing noises, and an external transformer to capture long-range feature dependencies. As opposed to existing architectures which adopt fixed and limited receptive fields, our method benefits from size-adaptive receptive fields as well as global contextual information, and outperforms current state-of-the-arts with significant improvements on popular datasets.

cs.CV↗
arXiv2022

UPST-NeRF: Universal Photorealistic Style Transfer of Neural Radiance Fields for 3D Scene

3D scenes photorealistic stylization aims to generate photorealistic images from arbitrary novel views according to a given style image while ensuring consistency when rendering from different viewpoints. Some existing stylization methods with neural radiance fields can effectively predict stylized scenes by combining the features of the style image with multi-view images to train 3D scenes. However, these methods generate novel view images that contain objectionable artifacts. Besides, they cannot achieve universal photorealistic stylization for a 3D scene. Therefore, a styling image must retrain a 3D scene representation network based on a neural radiation field. We propose a novel 3D scene photorealistic style transfer framework to address these issues. It can realize photorealistic 3D scene style transfer with a 2D style image. We first pre-trained a 2D photorealistic style transfer network, which can meet the photorealistic style transfer between any given content image and style image. Then, we use voxel features to optimize a 3D scene and get the geometric representation of the scene. Finally, we jointly optimize a hyper network to realize the scene photorealistic style transfer of arbitrary style images. In the transfer stage, we use a pre-trained 2D photorealistic network to constrain the photorealistic style of different views and different style images in the 3D scene. The experimental results show that our method not only realizes the 3D photorealistic style transfer of arbitrary style images but also outperforms the existing methods in terms of visual quality and consistency. Project page:https://semchan.github.io/UPST_NeRF.

cs.CV↗
arXiv2022

Theoretical study of small signal modulation behavior of Fabry-Perot Germanium-on-Silicon lasers

This work investigated the small signal performance of Fabry-Perot Ge-on-Si lasers by modeling and simulations. The 3dB bandwidth dependence on the structure parameters such as poly-Si cladding thickness, Ge cavity width and thickness, and minority carrier lifetime were studied. A 3dB bandwidth of 33.94 GHz at a biasing current of 270.5 mA is predicted after Ge laser structure optimization with a defect limited carrier lifetime of 1 ns.

physics.optics↗
arXiv2022

Sub-additive unstable topological pressure on saturated subsets

In this paper, we continue our investigation on sub-additive pressures for $C^1$-smooth partially hyperbolic diffeomorphisms. Under the assumption of unstable almost product property, we show that the unstable Bowen topological pressure on the saturated set of a given non-empty compact connected set of invariant measures equals the infimum of the summation of unstable metric entropy and the corresponding \emph{Lyapunov exponent}, where the infimum is taken over all invariant measures inside the compact connected set above. Moreover, we also show that the unstable topological capacity pressure on the saturated sets above coincide with the unstable topological pressure of the whole system.

math.DS↗
arXiv2021

Local Optical Conductivity of Bilayer Graphene with Kink Potential

We study the optical response of bilayer graphene with a kink potential composed of a domain wall separating two AB regions with opposite interlayer bias. The band structure and the local optical conductivity in real space are investigated in details based on a continuum model. We find that the one-dimensional chiral states localized at the domain wall contribute significantly to the local optical conductivity, which shows a clear distinction in different regions. The effects of domain wall states on optical conductivity can be detected by spatially and frequency resolved spectroscopic features. From the spectrum at various Fermi energies, important features in the band structure such as the energy separation between two chiral states can be directly measured. When the domain wall region is broad, the spatial distribution of local optical conductivity can provide important information on the bound states as well as the topological domain wall states.

cond-mat.mes-hall↗
arXiv2021

Effects of heterostrain and lattice relaxation on optical conductivity of twisted bilayer graphene

We present a theoretical study of the effects of heterostrain and lattice relaxation on the optical conductivity of twisted bilayer graphene near the magic angle, based on the band structures obtained from a continuum model. We find that heterostrain, lattice relaxation and their combination give rise to very distinctive spectroscopic features in the optical conductivity, which can be used to probe and distinguish these effects. From the spectrum at various Fermi energies, important features in the strain- and relaxation-modified band structure such as the bandgap, bandwidth and van Hove singularities can be directly measured. The peak associated with the transition between the flat bands in the optical conductivity are highly sensitive to the direction of the strain, which can provide direct information on the strain-modified flat bands.

cond-mat.mes-hall↗
arXiv2020

Optical Conductivity of Twisted Bilayer Graphene Near the Magic Angle

We theoretically study the band structure and optical conductivity of twisted bilayer graphene (TBG) near the magic angle considering the effects of lattice relaxation. We show that the optical conductivity spectrum is characterized by a series of peaks associated with the van Hove singularities in the band structure, and the peak energies evolve systematically with the twist angle. Lattice relaxation effects in TBG modify its band structure, especially the flat bands, which lead to significant shifts of the peaks in optical conductivity. These results demonstrate that spectroscopic features in the optical conductivity can serve as fingerprints for exploring the band structure, band gap and lattice relaxation in magic angle TBG as well as identifying its rotation angle.

cond-mat.mes-hall↗
arXiv2020

Unstable Topological Pressure for Partially Hyperbolic Diffeomorphisms with Sub-additive Potentials

In this paper, we introduce the unstable topological pressure for C^1-smooth partially hyperbolic diffeomorphisms with sub-additive potentials. Moreover, without any additional assumption, we have established the expected variational principle which connects this unstable topological pressure and the unstable measure theoretic entropy, as well as the corresponding Lyapunov exponent.

math.DS↗
arXiv2020

Deep Residual Dense U-Net for Resolution Enhancement in Accelerated MRI Acquisition

Typical Magnetic Resonance Imaging (MRI) scan may take 20 to 60 minutes. Reducing MRI scan time is beneficial for both patient experience and cost considerations. Accelerated MRI scan may be achieved by acquiring less amount of k-space data (down-sampling in the k-space). However, this leads to lower resolution and aliasing artifacts for the reconstructed images. There are many existing approaches for attempting to reconstruct high-quality images from down-sampled k-space data, with varying complexity and performance. In recent years, deep-learning approaches have been proposed for this task, and promising results have been reported. Still, the problem remains challenging especially because of the high fidelity requirement in most medical applications employing reconstructed MRI images. In this work, we propose a deep-learning approach, aiming at reconstructing high-quality images from accelerated MRI acquisition. Specifically, we use Convolutional Neural Network (CNN) to learn the differences between the aliased images and the original images, employing a U-Net-like architecture. Further, a micro-architecture termed Residual Dense Block (RDB) is introduced for learning a better feature representation than the plain U-Net. Considering the peculiarity of the down-sampled k-space data, we introduce a new term to the loss function in learning, which effectively employs the given k-space data during training to provide additional regularization on the update of the network weights. To evaluate the proposed approach, we compare it with other state-of-the-art methods. In both visual inspection and evaluation using standard metrics, the proposed approach is able to deliver improved performance, demonstrating its potential for providing an effective solution.

eess.IV↗
arXiv2019

Modelling and Dynamic Tracking Control of Industrial Vehicles with Tractor-trailer Structure

Existing works on control of tractor-trailers systems only consider the kinematics model without taking dynamics into account. Also, most of them treat the issue as a pure control theory problem whose solutions are difficult to implement. This paper presents a trajectory tracking control approach for a full-scale industrial tractor-trailers vehicle composed of a car-like tractor and arbitrary number of passive full trailers. To deal with dynamic effects of trailing units, a force sensor is innovatively installed at the connection between the tractor and the first trailer to measure the forces acting on the tractor. The tractor's dynamic model that explicitly accounts for the measured forces is derived. A tracking controller that compensates the pulling/pushing forces in real time and simultaneously drives the system onto desired trajectories is proposed. The propulsion map between throttle opening and the propulsion force is proposed to be modeled with a fifth-order polynomial. The parameters are estimated by fitting experimental data, in order to provide accurate driving force. Stability of the control algorithm is rigorously proved by Lyapunov methods. Experiments of full-size vehicles are conducted to validate the performance of the control approach.

eess.SY↗
arXiv2019

Unstable Metric Pressure of Partially Hyperbolic Diffeomorphisms with sub-additive Potentials

In this paper, we define and study unstable measure theoretic pressure for C^1-smooth partially hyperbolic diffeomorphisms with sub-additive potentials. For any ergodic measure, we show that this unstable metric pressure equals the corresponding unstable measure theoretic entropy plus the Lyapunov exponents with respect to the ergodic measure. On the other hand, we also define unstable topological and metric pressure in terms of the Bowen's picture and the capacity picture. We show that all these definitions of unstable metric pressure actually coincide for any ergodic measure.

math.DS↗
arXiv2018

Thomsen-Li's Theorem Revisited

In this paper, we prove a general version of Thomsen-Li's Theorem--a Krein-Milman type theorem for C*-algebras. To be precise, for a Markov operator on $C[0,1]$ which preserves certain subspace of $C[0,1]$, we approximate it by an average of homomorphisms on $C[0,1]$ in the strong operator topology, additionally we require that this average also preserves the same subspace. Such results could be useful for constructing homomorphisms for subhomogeneous C*-algebras.

math.OA↗
arXiv2018

Verma modules for rank two Heisenberg-Virasoro algebra

Let $\preceq$ be a compatible total order on the additive group $\mathbb{Z}^2$, and $L$ be the rank two Heisenberg-Virasoro algebra. For any $\mathbf{c}=(c_1,c_2,c_3,c_4) \in \mathbb{C}^4$, we define $\mathbb{Z}^2$-graded Verma module $M(\mathbf{c}, \preceq)$ for the Lie algebra $L$. A necessary and sufficient condition for the Verma module $M(\mathbf{c}, \preceq)$ to be irreducible is provided. Moreover, the maximal $\mathbb{Z}^2$-graded submodules of the Verma module $M(\mathbf{c}, \preceq)$ are characterized when $M(\mathbf{c}, \preceq)$ is reducible.

math.RT↗
arXiv2018

Optical Signatures of Dirac Nodal-lines in NbAs$_2$

Using polarized optical and magneto-optical spectroscopy, we have demonstrated universal aspects of electrodynamics associated with Dirac nodal-lines. We investigated anisotropic electrodynamics of NbAs$_2$ where the spin-orbit interaction triggers energy gaps along the nodal-lines, which manifest as sharp steps in the optical conductivity spectra. We show experimentally and theoretically that shifted 2D Dirac nodal-lines feature linear scaling $σ_1 (ω)\simω$, similar to 3D nodal-points. Massive Dirac nature of the nodal-lines are confirmed by magneto-optical data, which may also be indicative of theoretically predicted surface states. Optical data also offer a natural explanation for the giant magneto-resistance in NbAs$_2$.

cond-mat.mtrl-sci↗
arXiv2018

Prime orbit theorems for expanding Thurston maps

We obtain an analogue of the prime number theorem for a class of branched covering maps on the $2$-sphere called expanding Thurston maps $f$, which are topological models of some rational maps without any smoothness or holomorphicity assumption. More precisely, by studying dynamical zeta functions and, more generally, dynamical Dirichlet series for $f$, we show that the number of primitive periodic orbits of $f$, ordered by a weight on each point induced by a non-constant (eventually) positive real-valued Hölder continuous function $ϕ$ on $S^2$ satisfying some additional regularity conditions, is asymptotically the same as the well-known logarithmic integral, with an exponential error term. Such a result, known as a Prime Orbit Theorem, follows from our quantitative study of the holomorphic extension properties of the associated dynamical zeta functions and dynamical Dirichlet series. In particular, the above result applies to postcritically-finite rational maps whose Julia set is the whole Riemann sphere. Moreover, we prove that the regularity conditions needed here are generic; and for a Lattès map $f$ and a continuously differentiable (eventually) positive function $ϕ$, such a Prime Orbit Theorem holds if and only if $ϕ$ is not co-homologous to a constant.

math.DS↗
arXiv2018

On Harish-Chandra modules of the Lie algebra arising from the $2$-Dimensional Torus

Let $A=\mathbb{C}[t_1^{\pm1},t_2^{\pm1}]$ be the algebra of Laurent polynomials in two variables and $B$ be the set of skew derivations of $A$. Let $L$ be the universal central extension of the derived Lie subalgebra of the Lie algebra $A\rtimes B$. Set $\widetilde{L}=L\oplus\mathbb{C} d_1\oplus\mathbb{C} d_2$, where $d_1$, $d_2$ are two degree derivations. A Harish-Chandra module is defined as an irreducible weight module with finite dimensional weight spaces. In this paper, we prove that a Harish-Chandra module of the Lie algebra $\widetilde{L}$ is a uniformly bounded module or a generalized highest weight (GHW for short) module. Furthermore, we prove that the nonzero level Harish-Chandra modules of $\widetilde{L}$ are GHW modules. Finally, we classify all the GHW Harish-Chandra modules of $\widetilde{L}$.

math.RT↗