arXiv subjects

Dong Zhang

Publications and source records attributed to Dong Zhang.

At least 217 records · Page 12Linked to original sources

arXiv2018

Numerical simulations of supernova remnants in turbulent molecular clouds

Core-collapse supernova (SN) explosions may occur in the highly inhomogeneous molecular clouds (MCs) in which their progenitors were born. We perform a series of 3-dimensional hydrodynamic simulations to model the interaction between an individual supernova remnant (SNR) and a turbulent MC medium, in order to investigate possible observational evidence for the turbulent structure of MCs. We find that the properties of SNRs are mainly controlled by the mean density of the surrounding medium, while a SNR in a more turbulent medium with higher supersonic turbulent Mach number shows lower interior temperature, lower radial momentum, and dimmer X-ray emission compared to one in a less turbulent medium with the same mean density. We compare our simulations to observed SNRs, in particular, to W44, W28 and IC 443. We estimate that the mean density of the ambient medium is $\sim 10\,$cm$^{-3}$ for W44 and W28. The MC in front of IC 443 has a density of $\sim 100\,$cm$^{-3}$. We also predict that the ambient MC of W44 is more turbulent than that of W28 and IC 443. The ambient medium of W44 and W28 has significantly lower average density than that of the host giant MC. This result may be related to the stellar feedback from the SNRs' progenitors. Alternatively, SNe may occur close to the interface between molecular gas and lower density atomic gas. The region of shocked MC is then relatively small and the breakout into the low density atomic gas comprises most of the SNR volume.

astro-ph.HE↗

arXiv2018

Non-crossing chords of a polygon with forbidden positions

In this paper, we systematically study non-crossing chords of simple polygons in the plane. We first introduce the reduced Euler characteristic of a family of line-segments, and subsequently investigate the structure of the diagonals and epigonals of a polygon. Interestingly enough, the reduced Euler characteristic of a subfamily of diagonals and epigonals characterizes the geometric convexity of polygons. In particular, an alternative and complete answer is given for a problem proposed by G. C. Shephard. Meanwhile, we extend such research to non-crossing diagonals and epigonals with forbidden positions in some appropriate sense. We prove that the reduced Euler characteristic of diagonals with forbidden positions only depends on the information involving convex partitions by those forbidden diagonals, and it determines the shapes of polygons in a surprising way. Incidentally, some kinds of generalized Catalan's numbers naturally arise.

math.CO↗

arXiv2018

Connected door spaces and topological solutions of equations

The connected door space is an enigmatic topological space in which every proper nonempty subset is either open or closed, but not both. This paper provides an elementary proof of the classification theorem of connected door spaces. More importantly, we show that connected door topologies can be viewed as solutions of the valuation $f(A)+f(B)=f(A\cup B)+f(A\cap B)$ and the equation $f(A)+f(B)=f(A\cup B)$, respectively. In addition, some special solutions, which can be regarded as a union of connected door spaces, are provided.

math.GN↗

arXiv2018

Composition Loss for Counting, Density Map Estimation and Localization in Dense Crowds

With multiple crowd gatherings of millions of people every year in events ranging from pilgrimages to protests, concerts to marathons, and festivals to funerals; visual crowd analysis is emerging as a new frontier in computer vision. In particular, counting in highly dense crowds is a challenging problem with far-reaching applicability in crowd safety and management, as well as gauging political significance of protests and demonstrations. In this paper, we propose a novel approach that simultaneously solves the problems of counting, density map estimation and localization of people in a given dense crowd image. Our formulation is based on an important observation that the three problems are inherently related to each other making the loss function for optimizing a deep CNN decomposable. Since localization requires high-quality images and annotations, we introduce UCF-QNRF dataset that overcomes the shortcomings of previous datasets, and contains 1.25 million humans manually marked with dot annotations. Finally, we present evaluation measures and comparison with recent deep CNN networks, including those developed specifically for crowd counting. Our approach significantly outperforms state-of-the-art on the new dataset, which is the most challenging dataset with the largest number of crowd annotations in the most diverse set of scenes.

cs.CV↗

arXiv2018

Topological-charge dependence of radiation torque for an acoustic-vortex spanner

Based on the analyses of the wave front and the wave vector of an acoustic-vortex (AV) spanner generated by a circular array of point source, the principle of object rotation is investigated through the calculation of the orbital angular momentum (OAM) and the radiation torque for AVs with various topological charges. It is demonstrated theoretically that the rotation of the axisymmetric disk centered on an AV spanner is mainly driven by the exerted radiation torque without the contribution of the OAM transfer. The radiation torque on a small-radius object is inversely associated with the topological charge in the center of the AV spanner, and it is enhanced significantly for a larger AV with a higher topological charge. The special case of the radiation torque proportional to the topological charge might be realized when the disk radius is much larger than the wavelength, in which case the acoustic power of the AV spanner can be absorbed as much as possible. With the established sixteen-source experimental setup, the radial pressure distributions of AVs with different topological charges measured at the frequency of 1.3 kHz in air agree well with the simulations. The topological-charge dependence of the radiation torque for AVs is also verified by the quantitative laser-displacement (angle) measurements for disks with different radii. The favorable results demonstrate that, for the object covering the vortex center of an AV spanner, the total OAM transfer might be 0 and the object rotation is contributed by the acting moments. Thus, the radiation torque of an AV spanner is more applicable than the OAM for describing the driving capability of object rotation, and it can be used as an effective tool in clinical applications to noninvasively manipulate objects with a feature size at the wavelength-scale (e.g. kidney stone in lithotripsy) inside body.

physics.app-ph↗

arXiv2018

Dusty Cloud Acceleration by Radiation Pressure in Rapidly Star-Forming Galaxies

We perform two-dimensional and three-dimensional radiation hydrodynamic simulations to study cold clouds accelerated by radiation pressure on dust in the environment of rapidly star-forming galaxies dominated by infrared flux. We utilize the reduced speed of light approximation to solve the frequency-averaged, time-dependent radiative transfer equation. We find that radiation pressure is capable of accelerating the clouds to hundreds of kilometers per second while remaining dense and cold, consistent with observations. We compare these results to simulations where acceleration is provided by entrainment in a hot wind, where the momentum injection of the hot flow is comparable to the momentum in the radiation field. We find that the survival time of the cloud accelerated by the radiation field is significantly longer than that of a cloud entrained in a hot outflow. We show that the dynamics of the irradiated cloud depends on the initial optical depth, temperature of the cloud, and the intensity of the flux. Additionally, gas pressure from the background may limit cloud acceleration if the density ratio between the cloud and background is $\lesssim 10^{2}$. In general, a 10 pc-scale optically thin cloud forms a pancake structure elongated perpendicular to the direction of motion, while optically thick clouds form a filamentary structure elongated parallel to the direction of motion. The details of accelerated cloud morphology and geometry can also be affected by other factors, such as the cloud lengthscale, the reduced speed of light approximation, spatial resolution, initial cloud structure, and the dimensionality of the run, but these have relatively little affect on the cloud velocity or survival time.

astro-ph.GA↗

arXiv2017

ClusterNet: Detecting Small Objects in Large Scenes by Exploiting Spatio-Temporal Information

Object detection in wide area motion imagery (WAMI) has drawn the attention of the computer vision research community for a number of years. WAMI proposes a number of unique challenges including extremely small object sizes, both sparse and densely-packed objects, and extremely large search spaces (large video frames). Nearly all state-of-the-art methods in WAMI object detection report that appearance-based classifiers fail in this challenging data and instead rely almost entirely on motion information in the form of background subtraction or frame-differencing. In this work, we experimentally verify the failure of appearance-based classifiers in WAMI, such as Faster R-CNN and a heatmap-based fully convolutional neural network (CNN), and propose a novel two-stage spatio-temporal CNN which effectively and efficiently combines both appearance and motion information to significantly surpass the state-of-the-art in WAMI object detection. To reduce the large search space, the first stage (ClusterNet) takes in a set of extremely large video frames, combines the motion and appearance information within the convolutional architecture, and proposes regions of objects of interest (ROOBI). These ROOBI can contain from one to clusters of several hundred objects due to the large video frame size and varying object density in WAMI. The second stage (FoveaNet) then estimates the centroid location of all objects in that given ROOBI simultaneously via heatmap estimation. The proposed method exceeds state-of-the-art results on the WPAFB 2009 dataset by 5-16% for moving objects and nearly 50% for stopped objects, as well as being the first proposed method in wide area motion imagery to detect completely stationary objects.

cs.CV↗

arXiv2017

Coexistence of New Fermions in Topological Semimetal TaS

We theoretically propose that, the single crystal formed TaS is a new type of topological semimetal, hosting ring-shaped gapless nodal lines and triply degenerate points (TDPs) in the absence of spin-orbit coupling (SOC). In the presence of SOC, the each TDP splits into four TDPs along the high symmetric line in the momentum space, and one of the nodal ring remains closed due to the protection of the mirror reflection symmetry, while another nodal ring is fully gapped and transforms into six pairs ofWeyl points (WPs) carrying opposite chirality. The electronic structures of the projected surfaces are also discussed, the unique Fermi arcs are observed and the chirality remains or vanishes depending on the projection directions. On the (010) projected surface, one may observe a Lifshitz transition. The new type of topological semimetal TaS is stable and experimentally achievable, and the coexistence of topological nodal lines, WPs and TDPs states in TaS makes it a potential candidate to study the interplay between different types of topological fermions.

cond-mat.mes-hall↗

arXiv2017

Video Fill In the Blank using LR/RL LSTMs with Spatial-Temporal Attentions

Given a video and a description sentence with one missing word (we call it the "source sentence"), Video-Fill-In-the-Blank (VFIB) problem is to find the missing word automatically. The contextual information of the sentence, as well as visual cues from the video, are important to infer the missing word accurately. Since the source sentence is broken into two fragments: the sentence's left fragment (before the blank) and the sentence's right fragment (after the blank), traditional Recurrent Neural Networks cannot encode this structure accurately because of many possible variations of the missing word in terms of the location and type of the word in the source sentence. For example, a missing word can be the first word or be in the middle of the sentence and it can be a verb or an adjective. In this paper, we propose a framework to tackle the textual encoding: Two separate LSTMs (the LR and RL LSTMs) are employed to encode the left and right sentence fragments and a novel structure is introduced to combine each fragment with an "external memory" corresponding the opposite fragments. For the visual encoding, end-to-end spatial and temporal attention models are employed to select discriminative visual representations to find the missing word. In the experiments, we demonstrate the superior performance of the proposed method on challenging VFIB problem. Furthermore, we introduce an extended and more generalized version of VFIB, which is not limited to a single blank. Our experiments indicate the generalization capability of our method in dealing with such more realistic scenarios.

cs.CV↗

arXiv2017

Unsupervised Action Proposal Ranking through Proposal Recombination

Recently, action proposal methods have played an important role in action recognition tasks, as they reduce the search space dramatically. Most unsupervised action proposal methods tend to generate hundreds of action proposals which include many noisy, inconsistent, and unranked action proposals, while supervised action proposal methods take advantage of predefined object detectors (e.g., human detector) to refine and score the action proposals, but they require thousands of manual annotations to train. Given the action proposals in a video, the goal of the proposed work is to generate a few better action proposals that are ranked properly. In our approach, we first divide action proposal into sub-proposal and then use Dynamic Programming based graph optimization scheme to select the optimal combinations of sub-proposals from different proposals and assign each new proposal a score. We propose a new unsupervised image-based actioness detector that leverages web images and employs it as one of the node scores in our graph formulation. Moreover, we capture motion information by estimating the number of motion contours within each action proposal patch. The proposed method is an unsupervised method that neither needs bounding box annotations nor video level labels, which is desirable with the current explosion of large-scale action datasets. Our approach is generic and does not depend on a specific action proposal method. We evaluate our approach on several publicly available trimmed and un-trimmed datasets and obtain better performance compared to several proposal ranking methods. In addition, we demonstrate that properly ranked proposals produce significantly better action detection as compared to state-of-the-art proposal based methods.

cs.CV↗

arXiv2017

Entrainment in Trouble: Cool Cloud Acceleration and Destruction in Hot Supernova-Driven Galactic Winds

Efficient thermalization of overlapping supernovae within star-forming galaxies may produce a supernova-heated fluid that drives galactic winds. For fiducial assumptions about the timescale for cloud shredding from high-resolution simulations (which neglect magnetic fields) we show that cool clouds with temperature from $T_{c}\sim 10^{2}-10^{4}$ K seen in emission and absorption in galactic winds cannot be accelerated to observed velocities by the ram pressure of a hot wind. Taking into account both the radial structure of the hot flow and gravity, we show that this conclusion holds over a wide range of galaxy, cloud, and hot wind properties. This finding calls into question the prevailing picture whereby the cool atomic gas seen in galactic winds is entrained and accelerated by the hot flow. Given these difficulties with ram pressure acceleration, we discuss alternative models for the origin of high velocity cool gas outflows. Another possibility is that magnetic fields in cool clouds are sufficiently important that they prolong the cloud's life. For $T_{c}=10^{3}$\,K and $10^{4}$\,K clouds, we show that if conductive evaporation can be neglected, the cloud shredding timescale must be $\sim15$ and 5 times longer, respectively, than the values from hydrodynamical simulations in order for cool cloud velocities to reach those seen in observations.

astro-ph.GA↗

arXiv2017

Radiation Hydrodynamic Simulations of Dust-Driven Winds

We study dusty winds driven by radiation pressure in the atmosphere of a rapidly star-forming environment. We apply the variable Eddington tensor algorithm to re-examine the two-dimensional radiation hydrodynamic problem of a column of gas that is accelerated by a constant infrared radiation flux. In the absence of gravity, the system is primarily characterized by the initial optical depth of the gas. We perform several runs with different initial optical depth and resolution. We find that the gas spreads out along the vertical direction, as its mean velocity and velocity dispersion increase. In contrast to previous work using flux-limited diffusion algorithm, we find little evolution in the trapping factor. The momentum coupling between radiation and gas in the absence of gravity is similar to that with gravity. For Eddington ratio increasing with the height in the system, the momentum transfer from the radiation to the gas is not merely $\sim L/c$, but amplified by a factor of $1+ητ_{\rm IR}$, where $τ_{\rm IR}$ is the integrated infrared optical depth through the system, and $η\sim0.5-0.9$, decreasing with the optical depth. We apply our results to the atmosphere of galaxies and conclude that radiation pressure may be an important mechanism for driving winds in the most rapidly star-forming galaxies and starbursts.

astro-ph.GA↗

arXiv2017

Fast multicolor photodetectors based on graphene-contacted p-GaSe/n-InSe van der Waals heterostructures

The integration of different two-dimensional materials within a multilayer van der Waals (vdW) heterostructure offers a promising technology for realizing high performance opto-electronic devices such as photodetectors and light sources1-3. Transition metal dichalcogenides, e.g. MoS2 and WSe2, have been employed as the optically-active layer in recently developed heterojunctions. However, MoS2 and WSe2 become direct band gap semiconductors only in mono- or bilayer form4,5. In contrast, the metal monochalcogenides InSe and GaSe retain a direct bandgap over a wide range of layer thicknesses from bulk crystals down to exfoliated flakes only a few atomic monolayers thick6,7. Here we report on vdW heterojunction diodes based on InSe and GaSe: the type II band alignment between the two materials and their distinctive spectral response, combined with the low electrical resistance of transparent graphene electrodes, enable effective separation and extraction of photoexcited carriers from the heterostructure even when no external voltage is applied. Our devices are fast (< 10 μs), self-driven photodetectors with multicolor photoresponse ranging from the ultraviolet to the near-infrared and have the potential to accelerate the exploitation of two-dimensional vdW crystals by creating new routes to miniaturized optoelectronics beyond present technologies.

physics.ins-det↗

arXiv2016

Towards Fast Multicolor Photodetectors based on Graphene Contacted Vertical p-GaSe/n-InSe van der Waals Heterostructure

We investigated the electronic and optoelectronic properties of vertical van der Waals heterostructure photodetectors using layered p type GaSe and n type InSe, with graphene as the transparent electrodes. Not only the photocurrent peaks from the layered GaSe and InSe themselves were observed, also the interlayer optical transition peak was observed, which is consistent with the first-principles calculation. The built-in electric field between p-n heterojunction and the advantage of the graphene electrodes can effectively separate the photo-induced electron-hole pairs, and thus lead to the response time down to 160 μs. Making use of the interlayer as well as intralayer optical transitions of the vertical layered p-n heterostructure and graphene electrodes, we could achieve high performance multi-color photodetectors based on two-dimensional layered materials, which will be important for future optoelectronics.

cond-mat.mes-hall↗

arXiv2016

Video Fill in the Blank with Merging LSTMs

Given a video and its incomplete textural description with missing words, the Video-Fill-in-the-Blank (ViFitB) task is to automatically find the missing word. The contextual information of the sentences are important to infer the missing words; the visual cues are even more crucial to get a more accurate inference. In this paper, we presents a new method which intuitively takes advantage of the structure of the sentences and employs merging LSTMs (to merge two LSTMs) to tackle the problem with embedded textural and visual cues. In the experiments, we have demonstrated the superior performance of the proposed method on the challenging "Movie Fill-in-the-Blank" dataset.

cs.CV↗

arXiv2016

Topological Dirac semimetal phases in InSb/$α$-Sn semiconductor superlattices

We demonstrate theoretically the coexistence of Dirac semimetal and topological insulator phases in InSb/$α$-Sn conventional semiconductor superlattices, based on advanced first-principles calculations combined with low-energy $k\cdot p$ theory. By proper interfaces designing, a large interface polarization emerges when the growth direction is chosen along {[}111{]}. Such an intrinsic polarized electrostatic field reduces band gap largely and invert the band structure finally, leading to emerge of the topological Dirac semimetal phase with a pair of Dirac nodes appearing along the (111) crystallographic direction near the $Γ$ point. The surface states and Fermi arc are clearly observed in (100) projected surface. In addition, we also find a two-dimensional topological insulator phase with large nontrivial band gap approaching 70 meV, which make it possible to observe the quantum spin Hall effect at room temperature. Our proposal paves a way to realize topological nontrivial phases coexisted in conventional semiconductor superlattices by proper interface designing.

cond-mat.mtrl-sci↗

arXiv2016

Local feature hierarchy for face recognition across pose and illumination

Even though face recognition in frontal view and normal lighting condition works very well, the performance degenerates sharply in extreme conditions. Recently there are many work dealing with pose and illumination problems, respectively. However both the lighting and pose variation will always be encountered at the same time. Accordingly we propose an end-to-end face recognition method to deal with pose and illumination simultaneously based on convolutional networks where the discriminative nonlinear features that are invariant to pose and illumination are extracted. Normally the global structure for images taken in different views is quite diverse. Therefore we propose to use the 1*1 convolutional kernel to extract the local features. Furthermore the parallel multi-stream multi-layer 1*1 convolution network is developed to extract multi-hierarchy features. In the experiments we obtained the average face recognition rate of 96.9% on multiPIE dataset,which improves the state-of-the-art of face recognition across poses and illumination by 7.5%. Especially for profile-wise positions, the average recognition rate of our proposed network is 97.8%, which increases the state-of-the-art recognition rate by 19%.

cs.CV↗

arXiv2016

A Framework for Human Pose Estimation in Videos

In this paper, we present a method to estimate a sequence of human poses in unconstrained videos. We aim to demonstrate that by using temporal information, the human pose estimation results can be improved over image based pose estimation methods. In contrast to the commonly employed graph optimization formulation, which is NP-hard and needs approximate solutions, we formulate this problem into a unified two stage tree-based optimization problem for which an efficient and exact solution exists. Although the proposed method finds an exact solution, it does not sacrifice the ability to model the spatial and temporal constraints between body parts in the frames; in fact it models the {\em symmetric} parts better than the existing methods. The proposed method is based on two main ideas: `Abstraction' and `Association' to enforce the intra- and inter-frame body part constraints without inducing extra computational complexity to the polynomial time solution. Using the idea of `Abstraction', a new concept of `abstract body part' is introduced to conceptually combine the symmetric body parts and model them in the tree based body part structure. Using the idea of `Association', the optimal tracklets are generated for each abstract body part, in order to enforce the spatiotemporal constraints between body parts in adjacent frames. A sequence of the best poses is inferred from the abstract body part tracklets through the tree-based optimization. Finally, the poses are refined by limb alignment and refinement schemes. We evaluated the proposed method on three publicly available video based human pose estimation datasets, and obtained dramatically improved performance compared to the state-of-the-art methods.

cs.CV↗