Searcharxiv⌕ Search

arXiv subjects

Lei Jin

Publications and source records attributed to Lei Jin.

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

QueryPose: Sparse Multi-Person Pose Regression via Spatial-Aware Part-Level Query

We propose a sparse end-to-end multi-person pose regression framework, termed QueryPose, which can directly predict multi-person keypoint sequences from the input image. The existing end-to-end methods rely on dense representations to preserve the spatial detail and structure for precise keypoint localization. However, the dense paradigm introduces complex and redundant post-processes during inference. In our framework, each human instance is encoded by several learnable spatial-aware part-level queries associated with an instance-level query. First, we propose the Spatial Part Embedding Generation Module (SPEGM) that considers the local spatial attention mechanism to generate several spatial-sensitive part embeddings, which contain spatial details and structural information for enhancing the part-level queries. Second, we introduce the Selective Iteration Module (SIM) to adaptively update the sparse part-level queries via the generated spatial-sensitive part embeddings stage-by-stage. Based on the two proposed modules, the part-level queries are able to fully encode the spatial details and structural information for precise keypoint regression. With the bipartite matching, QueryPose avoids the hand-designed post-processes and surpasses the existing dense end-to-end methods with 73.6 AP on MS COCO mini-val set and 72.7 AP on CrowdPose test set. Code is available at https://github.com/buptxyb666/QueryPose.

cs.CV↗
arXiv2022

Mean dimension of product spaces: a fundamental formula

Mean dimension is a topological invariant of dynamical systems, which originates with Mikhail Gromov in 1999 and which was studied with deep applications around 2000 by Elon Lindenstrauss and Benjamin Weiss within the framework of amenable group actions. Let a countable discrete amenable group $G$ act continuously on compact metrizable spaces $X$ and $Y$. Consider the product action of $G$ on the product space $X\times Y$. The product inequality for mean dimension is well known: $\mathrm{mdim}(X\times Y,G)\le\mathrm{mdim}(X,G)+\mathrm{mdim}(Y,G)$, while it was unknown for a long time if the product inequality could be an equality. In 2019, Masaki Tsukamoto constructed the first example of two different continuous actions of $G$ on compact metrizable spaces $X$ and $Y$, respectively, such that the product inequality becomes strict. However, there is still one longstanding problem which remains open in this direction, asking if there exists a continuous action of $G$ on some compact metrizable space $X$ such that $\mathrm{mdim}(X\times X,G)<2\cdot\mathrm{mdim}(X,G)$. We solve this problem. Somewhat surprisingly, we prove, in contrast to (topological) dimension theory, a rather satisfactory theorem: If an infinite (countable discrete) amenable group $G$ acts continuously on a compact metrizable space $X$, then we have $\mathrm{mdim}(X^n,G)=n\cdot\mathrm{mdim}(X,G)$, for any positive integer $n$. Our product formula for mean dimension, together with the example and inequality (stated previously), eventually allows mean dimension of product actions to be fully understood.

math.DS↗
arXiv2022

AdaptivePose++: A Powerful Single-Stage Network for Multi-Person Pose Regression

Multi-person pose estimation generally follows top-down and bottom-up paradigms. Both of them use an extra stage ($\boldsymbol{e.g.,}$ human detection in top-down paradigm or grouping process in bottom-up paradigm) to build the relationship between the human instance and corresponding keypoints, thus leading to the high computation cost and redundant two-stage pipeline. To address the above issue, we propose to represent the human parts as adaptive points and introduce a fine-grained body representation method. The novel body representation is able to sufficiently encode the diverse pose information and effectively model the relationship between the human instance and corresponding keypoints in a single-forward pass. With the proposed body representation, we further deliver a compact single-stage multi-person pose regression network, termed as AdaptivePose. During inference, our proposed network only needs a single-step decode operation to form the multi-person pose without complex post-processes and refinements. We employ AdaptivePose for both 2D/3D multi-person pose estimation tasks to verify the effectiveness of AdaptivePose. Without any bells and whistles, we achieve the most competitive performance on MS COCO and CrowdPose in terms of accuracy and speed. Furthermore, the outstanding performance on MuCo-3DHP and MuPoTS-3D further demonstrates the effectiveness and generalizability on 3D scenes. Code is available at https://github.com/buptxyb666/AdaptivePose.

cs.CV↗
arXiv2022

Mean dimension of Bernstein spaces and universal real flows

We study the action of translation on the spaces of uniformly bounded continuous functions on the real line which are uniformly band-limited in a compact interval. We prove that two intervals themselves will decide if two spaces are topologically conjugate, while the length of an interval tells the mean dimension of a space. We also investigate universal real flows. We construct a sequence of compact invariant subsets of a space consisting of uniformly bounded smooth one-Lipschitz functions on the real line, which have mean dimension equal to one, such that all real flows can be equivariantly embedded in the translation on their product space. Moreover, we show that the countable self-product of any among them does not satisfy such a universal property. This, on the one hand, presents a more reasonable choice of a universal real flow with a view towards mean dimension, and on the other hand, clarifies a seemingly plausible impression; meanwhile, it refines the previously known results in this direction. Our proof goes through an approach of harmonic analysis. Furthermore, both the universal space that we provide and an embedding mapping which we build for any real flow are explicit.

math.DS↗
arXiv2022

Strong Optomechanical Interactions with Long-lived Fundamental Acoustic Waves

Traveling-wave optomechanical interactions, known as Brillouin interactions, have now been established as a powerful and versatile resource for photonic sources, sensors, and radio-frequency processors. However, established Brillouin-based interactions with sufficient interaction strengths involve short phonon lifetimes, which critically limit their performance for applications including radio-frequency filtering and optomechanical storage devices. Here, we investigate a new paradigm of optomechanical interactions with fundamental acoustic modes, where interaction strength is decoupled from phonon lifetimes, enabling the uniquely desirable combination of high optomechanical coupling, long phonon lifetimes, tunable phonon frequencies, and single-sideband amplification. Using sensitive four-wave mixing spectroscopy controlling for noise and spatial mode coupling, optomechanical interactions with long > 2 $μs$ phonon lifetimes and strong > 400 $W^{-1} m^{-1}$ coupling are observed in a tapered fiber. In addition, we demonstrate novel phonon self-interference effects resulting from the unique combination of an axially varying device geometry with long phonon lifetimes. A generalized theoretical model, in excellent agreement with experiments, is developed with broad applicability to inhomogeneous optomechanical systems.

physics.optics↗
arXiv2022

Directional mean dimension and continuum-wise expansive $\mathbb{Z}^k$-actions

We study directional mean dimension of $\mathbb{Z}^k$-actions (where $k$ is a positive integer). On the one hand, we show that there is a $\mathbb{Z}^2$-action whose directional mean dimension (considered as a $[0,+\infty]$-valued function on the torus) is not continuous. On the other hand, we prove that if a $\mathbb{Z}^k$-action is continuum-wise expansive, then the values of its $(k-1)$-dimensional directional mean dimension are bounded. This is a generalization (with a view towards Meyerovitch and Tsukamoto's theorem on mean dimension and expansive multiparameter actions) of a classical result due to Mañé: Any compact metrizable space admitting an expansive homeomorphism (with respect to a compatible metric) is finite-dimensional.

math.DS↗
arXiv2022

Learning Quality-aware Representation for Multi-person Pose Regression

Off-the-shelf single-stage multi-person pose regression methods generally leverage the instance score (i.e., confidence of the instance localization) to indicate the pose quality for selecting the pose candidates. We consider that there are two gaps involved in existing paradigm:~1) The instance score is not well interrelated with the pose regression quality.~2) The instance feature representation, which is used for predicting the instance score, does not explicitly encode the structural pose information to predict the reasonable score that represents pose regression quality. To address the aforementioned issues, we propose to learn the pose regression quality-aware representation. Concretely, for the first gap, instead of using the previous instance confidence label (e.g., discrete {1,0} or Gaussian representation) to denote the position and confidence for person instance, we firstly introduce the Consistent Instance Representation (CIR) that unifies the pose regression quality score of instance and the confidence of background into a pixel-wise score map to calibrates the inconsistency between instance score and pose regression quality. To fill the second gap, we further present the Query Encoding Module (QEM) including the Keypoint Query Encoding (KQE) to encode the positional and semantic information for each keypoint and the Pose Query Encoding (PQE) which explicitly encodes the predicted structural pose information to better fit the Consistent Instance Representation (CIR). By using the proposed components, we significantly alleviate the above gaps. Our method outperforms previous single-stage regression-based even bottom-up methods and achieves the state-of-the-art result of 71.7 AP on MS COCO test-dev set.

cs.CV↗
arXiv2021

The 2nd Anti-UAV Workshop & Challenge: Methods and Results

The 2nd Anti-UAV Workshop \& Challenge aims to encourage research in developing novel and accurate methods for multi-scale object tracking. The Anti-UAV dataset used for the Anti-UAV Challenge has been publicly released. There are two subsets in the dataset, $i.e.$, the test-dev subset and test-challenge subset. Both subsets consist of 140 thermal infrared video sequences, spanning multiple occurrences of multi-scale UAVs. Around 24 participating teams from the globe competed in the 2nd Anti-UAV Challenge. In this paper, we provide a brief summary of the 2nd Anti-UAV Workshop \& Challenge including brief introductions to the top three methods.The submission leaderboard will be reopened for researchers that are interested in the Anti-UAV challenge. The benchmark dataset and other information can be found at: https://anti-uav.github.io/.

cs.CV↗
arXiv2021

Schrodinger's Red Pixel by Quasi Bound-State-In-Continuum

While structural colors are ubiquitous in nature, saturated reds are mysteriously absent. Hence, a longstanding problem is in fabricating nanostructured surfaces that exhibit reflectance approaching the theoretical limit. This limit is termed the Schrodinger red and demands sharp spectral transitions from "stopband" to a high reflectance "passband" with total suppression of higher-order resonances at blue and green wavelengths. Current approaches based on metallic or dielectric nanoantennas are insufficient to simultaneously meet these conditions. Here, for the 1st time, we designed and fabricated tall Si nanoantenna arrays on quartz substrate to support two partially overlapping y polarized quasi bound-state-in-the-continuum (q-BIC) modes in the red wavelengths with sharp spectral edges. These structures produce possibly the most saturated and brightest reds with ~80% reflectance, exceeding the red vertex in sRGB and even the cadmium red pigment. We employed a gradient descent algorithm with structures supporting q BIC as the starting point. Although the current design is polarization dependent, the proposed paradigm has enabled us to achieve the elusive structural red and the design principle could be generalized to Schrodinger's pixels of other colors. The design is suitable for scale up using other nanofabrication techniques for larger area applications, such as red pixels in displays, decorative coatings, and miniaturized spectrometers with high wavelength selectivity.

physics.optics↗
arXiv2021

Observation of unconventional six-fold, four-fold and three-fold excitations in rare-earth-metal carbide Re2C3

Unconventional fermions, such as three-fold, four-fold, six-fold, and eight-fold fermions have attracted intense attention in recent years. However, the concrete materials hosting unconventional fermions are still in urgent scarcity. In this work, based first-principle calculations and symmetry analysis, we reveal rich unconventional fermions in existing compound Re2C3 (Re = Y, La, Ce, Pr, Nd, Sm, Tb, Dy, Ho, Er, Tm, Yb, Lu). We show that these compounds host quadratic dispersive three-fold (TP), linear dispersive four-fold (FP) and six-fold points (SP) near the Fermi level in their electric band structures when spin-orbital coupling (SOC) is not included. Notably, the FP is charge-2 Dirac-like point. More importantly, among compound Re2C3, the compound Yb2C3 has very clean band structure, and its unconventional fermions are closed to the Fermi level. We also find that a uniaxial strain can transform the unconventional fermions into other types fermions, depending on the directions of strain. When SOC is considered, a SP transform to an eightfold degenerate point and a fourfold degenerate point. Overall, our work provides a family of realistic materials to study the unconventional fermions.

cond-mat.mtrl-sci↗
arXiv2020

SWCNT@BNNT with 1D van der Waals Heterostructure with a High Optical Damage Threshold for Laser Mode-locking

Single-walled carbon nanotube encapsulated in boron nitrite nanotube (SWCNT@BNNT) is a novel nanomaterial with a one-dimensional van der Waals (1D-vdW) heterostructure. In this paper, we demonstrated that the SWCNT@BNNT has an enhanced optical power tolerance compared to that of the pristine SWCNT while exhibiting an optical saturable absorption properties. Under optical power intensity of 13kW/cm2, the lifetime of the SWCNT@BNNT is found to be 2,270 times longer than SWCNT for 1% degradation in absorbance. A short-cavity mode-locked laser with a high repetition rate of 1 GHz has been realized using the SWCNT@BNNT as the saturable absorber. We have shown that the technique of fabricating nanomaterial with a 1D-vdW heterostructure can modify and enhance the optical properties of the encapsulated nanomaterials.

physics.optics↗
arXiv2020

Origin of the hump anomalies in the Hall resistance loops of ultrathin SrRuO$_3$/SrIrO$_3$ multilayers

The proposal that very small Néel skyrmions can form in SrRuO$_3$/SrIrO$_3$ epitaxial bilayers and that the electric field-effect can be used to manipulate these skyrmions in gated devices strongly stimulated the recent research of SrRuO$_3$ heterostructures. A strong interfacial Dzyaloshinskii-Moriya interaction, combined with the breaking of inversion symmetry, was considered as the driving force for the formation of skyrmions in SrRuO$_3$/SrIrO$_3$ bilayers. Here, we investigated nominally symmetric heterostructures in which an ultrathin ferromagnetic SrRuO$_3$ layer is sandwiched between large spin-orbit coupling SrIrO$_3$ layers, for which the conditions are not favorable for the emergence of a net interfacial Dzyaloshinskii-Moriya interaction. Previously the formation of skyrmions in the asymmetric SrRuO$_3$/SrIrO$_3$ bilayers was inferred from anomalous Hall resistance loops showing humplike features that resembled topological Hall effect contributions. Symmetric SrIrO$_3$/SrRuO$_3$/SrIrO$_3$ trilayers do not show hump anomalies in the Hall loops. However, the anomalous Hall resistance loops of symmetric multilayers, in which the trilayer is stacked several times, do exhibit the humplike structures, similar to the asymmetric SrRuO$_3$/SrIrO$_3$ bilayers. The origin of the Hall effect loop anomalies likely resides in unavoidable differences in the electronic and magnetic properties of the individual SrRuO$_3$ layers rather than in the formation of skyrmions.

cond-mat.mtrl-sci↗
arXiv2020

Fully spin-polarized double-Weyl fermions with type-III dispersion in quasi-one dimensional materials X2RhF6 (X=K, Rb, Cs)

Double-Weyl fermions, as novel topological states of matter, have been mostly discussed in nonmagnetic materials. Here, based on density-functional theory and symmetry analysis, we propose the realization of fully spin-polarized double-Weyl fermions in a family ferromagnetic materials X2RhF6 (X= K, Rb, Cs). These materials have the half-metal ground states, where only the bands from the spin-down channel present near the Fermi energy. The spin-down bands form a pair of triply degenerate nodal points (TDNPs) if spin-orbit coupling (SOC) is not included. Under SOC, one TDNP splits into two double-Weyl points featuring quadratic dispersion along two momentum direction, and they are protected by the three-fold rotation (C3) symmetry. Unlike most double-Weyl semimetals, the Weyl points proposed here have the type-III dispersion with one of the crossing bands being saddle-shaped. An effective model is constructed, which describes well the nature of the Weyl points. These Weyl points are fully spin-polarized, and are characterized with double Fermi arcs on the surface spectrum. Breaking C3 symmetry by lattice strain could shift one double-Weyl point into a pair of type-II single-Weyl points. The X2RhF6 materials proposed here are excellent candidates to investigate the novel properties of type-III double-Weyl fermions in ferromagnetic system, as well as generate potential applications in spintronics.

cond-mat.mtrl-sci↗
arXiv2020

P$^{2}$Net: Patch-match and Plane-regularization for Unsupervised Indoor Depth Estimation

This paper tackles the unsupervised depth estimation task in indoor environments. The task is extremely challenging because of the vast areas of non-texture regions in these scenes. These areas could overwhelm the optimization process in the commonly used unsupervised depth estimation framework proposed for outdoor environments. However, even when those regions are masked out, the performance is still unsatisfactory. In this paper, we argue that the poor performance suffers from the non-discriminative point-based matching. To this end, we propose P$^2$Net. We first extract points with large local gradients and adopt patches centered at each point as its representation. Multiview consistency loss is then defined over patches. This operation significantly improves the robustness of the network training. Furthermore, because those textureless regions in indoor scenes (e.g., wall, floor, roof, \etc) usually correspond to planar regions, we propose to leverage superpixels as a plane prior. We enforce the predicted depth to be well fitted by a plane within each superpixel. Extensive experiments on NYUv2 and ScanNet show that our P$^2$Net outperforms existing approaches by a large margin. Code is available at \url{https://github.com/svip-lab/Indoor-SfMLearner}.

cs.CV↗
arXiv2020

Nanoscale-femtosecond dielectric response of Mott insulators captured by two-colour near-field ultrafast electron microscopy

Characterizing and controlling the out-of-equilibrium state of nanostructured Mott insulators hold great promises for emerging quantum technologies while providing an exciting playground for investigating fundamental physics of strongly-correlated systems. Here, we use two-colour near-field ultrafast electron microscopy to photo-induce the insulator-to-metal transition in a single VO2 nanowire and probe the ensuing electronic dynamics with combined nanometer-femtosecond resolution. We take advantage of a femtosecond temporal gating of the electron pulse mediated by an infrared laser pulse, and exploit the sensitivity of inelastic electron-light scattering to changes in the material dielectric function. By spatially mapping the near-field dynamics of an individual nanowire of VO2, we observe that ultrafast photo-doping drives the system into a metallic state on a time scale of about 150 fs without yet perturbing the crystalline lattice. Due to the high versatility and sensitivity of the electron probe, our method would allow capturing the electronic dynamics of a wide range of nanoscale materials with ultimate spatio-temporal resolution.

physics.app-ph↗
arXiv2020

A Record-High Ion Storage Capacity of T-Graphene as Two-Dimensional Anode Material for Li-ion and Na-ion Batteries

Developing applicable two-dimensional (2D) electrode materials with high performance, especially with high ion storage capacity, has become an ever more obsessive quest in recent years. Based on first-principles calculations, we report that T-graphene, a new carbon-based 2D material, has a record-high Li/Na storage capacity. The capacity of T-graphene is as high as 2233.2 mA h g-1 for Li, and can reach 2357.2 mA h g-1 for Na, which are 6 times as much as that of the commercial graphite and are the highest among 2D anode materials identified so far. We demonstrate that the ultrahigh storage capacity of T-graphene mostly benefits from its low atomic mass and special periodic lattice structure. T-graphene has not only the ultrahigh storage capacity but also hosts the stable ion adsorption, good electric conductivity, fast ion diffusion speed, and low open-circuit voltage, which are merits required as a superior anode material for Li-ion and Na-ion batteries with ultrahigh storage capacity.

cond-mat.mtrl-sci↗
arXiv2020

Topological Nodal Line Electrides: Realization of Ideal Nodal Line State Nearly Immune from Spin-Orbit Coupling

Nodal line semimetals (NLSs) have attracted broad interest in current research. In most of existing NLSs, the intrinsic properties of nodal lines are greatly destroyed because nodal lines usually suffer sizable gaps induced by non-negligible spin-orbit coupling (SOC). In this work,we propose the topological nodal line electrides (TNLEs), which achieve electronic structures of nodal lines and electrides simultaneously, provide new insight on designing excellent NLSs nearly immune from SOC. Since the states near the Fermi level are most contributed by nonnucleus-bounded interstitial electrons, nodal lines in TNLEs manifest extremely small SOCinduced gap even possessing heavy elements. Especially, we propose the family of A2B (A = Ca, Sr, Ba; B= As, Sb, Bi) materials are realistic TNLEs with negligible SOC-induced gaps, which can play as excellent platforms to study the intrinsic properties of TNLEs

cond-mat.mtrl-sci↗